https://reunir.unir.net/handle/123456789/16201 2025-04-27T03:26:25Z 2025-04-27T03:26:25Z Mu-Yen, Chen C. K. Hung, Patrick https://reunir.unir.net/handle/123456789/17347 2024-09-04T07:42:11Z

Editor’s Note Mu-Yen, Chen; C. K. Hung, Patrick With (EC) the rise of global economy and Electronic Commerce (EC), efficient inter-organizational planning and deployment for value chain processes have become important. Radio-frequency Identification (RFID), Near Field Communication (NFC), and related wireless technologies are evaluated to be some of the most significant technological innovations in the twenty-first century. In the past few years, wireless and context-awareness technology have led to much hope and optimism. The mainstream press hails these innovations as the avant-garde in technology and business. The Internet of Everything (IoE) goal is the intelligent connection of people, process, data, and things. The IoE describes a world where billions of objects have sensors to detect, measure, and assess their status, all connected over public or private networks using standard and proprietary protocols. Hence, this special issue investigates the state-of-art AI and deep learning approaches for successful systems or applications in the IoE environment. In addition, this special issue also wants to understand the direct and indirect effects of using these smart technologies to build language information processing based on the Web of Things (WoT) around the smart cities and societies. Submitted by Eva María Arévalo Cid (evamaria.arevalo@unir.net) on 2024-09-04T07:42:11Z No. of bitstreams: 1 Editorial.pdf: 57324 bytes, checksum: 5223830d24ee8c02a8626180dc5aec60 (MD5); Made available in DSpace on 2024-09-04T07:42:11Z (GMT). No. of bitstreams: 1 Editorial.pdf: 57324 bytes, checksum: 5223830d24ee8c02a8626180dc5aec60 (MD5)

Acarer, Tayfun https://reunir.unir.net/handle/123456789/17346 2024-09-03T15:34:28Z

Energy-Aware Path Planning by Autonomous Underwater Vehicle in Underwater Wireless Sensor Networks for Safer Maritime Transportation Acarer, Tayfun Throughout history, maritime transportation has been preferred for international and intercontinental trade thanks to its lower cost than other transportation ways, which have a risk of ship accidents. To avoid these risks, underwater wireless sensor networks can be used as a robust and safe solution by monitoring maritime environment where energy resources are critical. Energy constraints must be solved to enable continuous data collection and communication for environmental monitoring and surveillance so they can last. Their energy limitations and battery replacement difficulties can be addressed with a path planning approach.This paper considers the energy-aware path planning problem with autonomous underwater vehicles by five commonly used approaches, namely, Ant Colony Optimization-based Approach, Particle Swarm Optimization-based Approach, Teaching Learning-based Optimization-based Approach, Genetic Algorithm-based Approach, Grey Wolf Optimizer-based Approach. Simulations show that the system converges faster and performs better with genetic algorithm than the others. This paper also considers the case where direct traveling paths between some node pairs should be avoided due to several reasons including underwater flows, too narrow places for travel, and other risks like changing temperature and pressure. To tackle this case, we propose a modified genetic algorithm, the Safety-Aware Genetic Algorithm-based Approach, that blocks the direct paths between those nodes. In this scenario, the Safety-Aware Genetic Algorithm-based approach provides just a 3% longer path than the Genetic Algorithm-based approach which is the best approach among all these approaches. This shows that the Safety-Aware Genetic Algorithm-based approach performs very robustly. With our proposed robust and energy-efficient path-planning algorithms, the data gathered by sensors can be collected very quickly with much less energy, which enables the monitoring system to respond faster for ship accidents. It also reduces total energy consumption of sensors by communicating them closely and so extends the network lifetime. Submitted by Eva María Arévalo Cid (evamaria.arevalo@unir.net) on 2024-09-03T15:34:28Z No. of bitstreams: 1 Energy-Aware Path Planning by Autonomous Underwater Vehicle in Underwater Wireless Sensor Networks for Safer Maritime Transportation.pdf: 1136398 bytes, checksum: c439f2ab4117c44a0ceb71f6494fcd27 (MD5); Made available in DSpace on 2024-09-03T15:34:28Z (GMT). No. of bitstreams: 1 Energy-Aware Path Planning by Autonomous Underwater Vehicle in Underwater Wireless Sensor Networks for Safer Maritime Transportation.pdf: 1136398 bytes, checksum: c439f2ab4117c44a0ceb71f6494fcd27 (MD5)

Wu, Cheng-Feng Zhang, Kunkun Lin, Meng-Chen Chiou, Chei-Chang https://reunir.unir.net/handle/123456789/17345 2024-09-03T15:20:07Z

Predicting Consumer Electronics E-Commerce: Technology Acceptance Model and Logistics Service Quality Wu, Cheng-Feng; Zhang, Kunkun; Lin, Meng-Chen; Chiou, Chei-Chang In online shopping for consumer electronics, information and physical flows are crucial determinants of consumer purchase intentions. This study examines these factors by integrating the Technology Acceptance Model with logistics service quality, analyzing the relationship between retailers and consumers in e-commerce. The focus is on how information and physical flows, as critical supply chain elements, affect consumers' decisions to purchase online. A structural model and machine learning algorithm with SHapley Additive exPlanations are employed to analyze the data, providing a comprehensive analysis of the Technology Acceptance Model in conjunction with logistics service quality. The findings reveal that attitude, perceived usefulness, and informativeness are the most influential factors affecting consumers' purchase intention. This study contributes to the understanding of consumer behavior in the context of e-commerce platforms for consumer electronic products by integrating the Technology Acceptance Model and logistics service quality theoretical perspectives and analyzing the data using innovative techniques, specifically, Shapley Additive Explanations. This research offers valuable insights into the significant role of various features in shaping consumers' purchase intention in the context of online e-commerce platforms for consumer electrical products. Submitted by Eva María Arévalo Cid (evamaria.arevalo@unir.net) on 2024-09-03T15:20:07Z No. of bitstreams: 1 Predicting Consumer Electronics E-Commerce Technology Acceptance Model and Logistics Service Quality.pdf: 693126 bytes, checksum: 87919cc7e704958addc421323ece5da9 (MD5); Made available in DSpace on 2024-09-03T15:20:07Z (GMT). No. of bitstreams: 1 Predicting Consumer Electronics E-Commerce Technology Acceptance Model and Logistics Service Quality.pdf: 693126 bytes, checksum: 87919cc7e704958addc421323ece5da9 (MD5)

Hongling, Wang https://reunir.unir.net/handle/123456789/17344 2024-09-03T15:01:47Z

Design of Traffic Electronic Information Signal Acquisition System Based on Internet of Things Technology and Artificial Intelligence Hongling, Wang This study aims to devise a traffic electronic information signal acquisition system employing Internet of Things and artificial intelligence technologies, offering a novel approach to address prevailing challenges related to traffic congestion and safety. Initially, the hardware circuit for the high-speed signal acquisition control core is developed, leveraging Field-Programmable Gate Array technology. This facilitates wireless monitoring of signal acquisition. Subsequently, a comprehensive time signal acquisition system is formulated, encompassing modules for communication, acquisition, storage, adaptive measurement, and signal analysis. The geomagnetic acquisition module within this system is utilized for collecting geomagnetic signals, which are then translated into switch signals indicating the presence or absence of vehicles. These signals are subsequently transmitted to the geomagnetic signal processor. Experimental results pertaining to the signal acquisition system reveal a notable peak storage speed of 200KB/s, considering the utilization of one million sampling points. Across a series of tests, the maximum relative error of the obtained results ranges from 2.2% to 2.7%, underscoring the consistency and reliability of the measurements. In comparison to existing testing devices, the system exhibits heightened accuracy in test results, rendering it more apt for traffic signal acquisition applications. In conclusion, this study accomplishes the collection and dissemination of diverse traffic information, furnishing robust support for traffic control and ensuring safe operations. Submitted by Eva María Arévalo Cid (evamaria.arevalo@unir.net) on 2024-09-03T15:01:46Z No. of bitstreams: 1 Design of Traffic Electronic Information Signal Acquisition System Based on Internet of Things Technology and Artificial Intelligence.pdf: 413839 bytes, checksum: 60a21bee29bd7ed62a03a51d3a79749c (MD5); Made available in DSpace on 2024-09-03T15:01:46Z (GMT). No. of bitstreams: 1 Design of Traffic Electronic Information Signal Acquisition System Based on Internet of Things Technology and Artificial Intelligence.pdf: 413839 bytes, checksum: 60a21bee29bd7ed62a03a51d3a79749c (MD5)

Agraz, Melih https://reunir.unir.net/handle/123456789/17202 2024-09-03T11:40:50Z

Semi-Supervised Machine Learning Approaches for Thyroid Disease Prediction and its Integration With the Internet of Everything Agraz, Melih Thyroid disorders are critical conditions that considerably affect a person’s general health, and may lead to additional health complications. Notably, these conditions often remain undetected in individuals who show "normal" results on traditional thyroid function tests. To enhance the diagnostic accuracy for thyroid disorders, such as hypothyroidism and hyperthyroidism, this study leveraged digital health records and explored semisupervised learning methods. We intentionally removed the labels from subjects initially categorized as "normal," incorporating them into our dataset as unlabeled data. The goal was to overcome the limitations of conventional diagnostic techniques, which may fail to detect subtle imbalances in thyroid hormones. In pursuit of this objective, we employed a combination of semi-supervised learning methods, namely FixMatch, Co-training, and self-training, in conjunction with supervised learning algorithms, specifically Naive Bayes and logistic regression. Our findings indicate that the FixMatch algorithm surpassed traditional supervised learning methods in various metrics, including accuracy (0.9054), sensitivity (0.9494), negative predictive value (0.9365), and F1 score (0.9146). Additionally, we propose a framework for integrating these diagnostic tools into the Internet of Everything (IoE) to promote early detection and facilitate improved healthcare outcomes. This research highlights the potential of semi-supervised learning techniques in the diagnosis of thyroid disorders and offers a roadmap for harnessing the IoE in healthcare advancement. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-08-07T15:34:48Z No. of bitstreams: 1 Semi-Supervised Machine Learning Approaches.pdf: 1416476 bytes, checksum: a99c331898f8c23044ac4ad09d770bf0 (MD5); Made available in DSpace on 2024-08-07T15:34:48Z (GMT). No. of bitstreams: 1 Semi-Supervised Machine Learning Approaches.pdf: 1416476 bytes, checksum: a99c331898f8c23044ac4ad09d770bf0 (MD5)

Yan, Lou Ren, Zhipeng Zhang, Yong Tao, Zhonghui Zhao, Yizu https://reunir.unir.net/handle/123456789/17201 2024-09-03T11:44:46Z

Constructing the Public Opinion Crisis Prediction Model Using CNN and LSTM Techniques Based on Social Network Mining Yan, Lou; Ren, Zhipeng; Zhang, Yong; Tao, Zhonghui; Zhao, Yizu This research endeavors to address the persistent dissemination of public opinion within social networks, mitigate the propagation of inappropriate content on these platforms, and enhance the overall service quality of social networks. To achieve these objectives, Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) techniques are employed in this research to develop a predictive model for anticipating public opinion crises in social network mining. This model furnishes users with a valuable reference for subsequent decisionmaking processes. The initial phase of this research involves the collection of user behavior data from social networks using IoT technologies, serving as the basis for extensive big data analysis and neural network research. Subsequently, a social network text categorization model is constructed by amalgamating the Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) architecture, elucidating the training procedures of deep learning methodologies within CNN and LSTM networks. The effectiveness of this approach is subsequently validated through comparisons with other deep learning techniques. Based on the obtained results and findings, the CNN-LSTM model demonstrates a noteworthy accuracy rate of 92.19% and an exceptionally low loss value of 0.4075. Of particular significance is the classification accuracy of the CNN-LSTM algorithm within social network datasets, which surpasses that of alternative algorithms, including CNN (by 6.31%), LSTM (by 4.43%), RNN (by 3.51%), Transformer (by 40.29%), and Generative Adversarial Network (GAN) (by 4.49%). This underscores the effectiveness of the CNN-LSTM algorithm in the realm of social network text classification. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-08-07T15:29:05Z No. of bitstreams: 1 Constructing the Public Opinion Crisis Prediction Model Using CNN and LSTM Techniques Based on Social Network Mining.pdf: 3327316 bytes, checksum: 4dc82685d20fce25bac3f02ec4bed882 (MD5); Made available in DSpace on 2024-08-07T15:29:05Z (GMT). No. of bitstreams: 1 Constructing the Public Opinion Crisis Prediction Model Using CNN and LSTM Techniques Based on Social Network Mining.pdf: 3327316 bytes, checksum: 4dc82685d20fce25bac3f02ec4bed882 (MD5)

Li, Hui Liu, Huayang Zhao, Wei Liu, Hao https://reunir.unir.net/handle/123456789/17200 2024-09-03T11:47:08Z

The Human Motion Behavior Recognition by Deep Learning Approach and the Internet of Things Li, Hui; Liu, Huayang; Zhao, Wei; Liu, Hao This paper is dedicated to exploring the practical implementation of deep learning and Internet of Things (IoT) technology within systems designed for recognizing human motion behavior. It places a particular emphasis on evaluating performance in complex environments, aiming to mitigate challenges such as poor robustness and high computational workload encountered in conventional human motion behavior recognition approaches by employing Convolutional Neural Networks (CNN). The primary focus is on enhancing the performance of human motion behavior recognition systems for real-world scenarios, optimizing them for real-time accuracy, and enhancing their suitability for practical applications. Specifically, the paper investigates human motion behavior recognition using CNN, where the parameters of the CNN model are fine-tuned to improve recognition performance. The paper commences by delineating the process and methodology employed for human motion recognition, followed by an in-depth exploration of the CNN model's application in recognizing human motion behavior. To acquire data depicting human motion behavior in authentic settings, the Internet of Things (IoT) is utilized for extracting relevant information from the living environment. The dataset chosen for human motion behavior recognition is the Royal Institute of Technology (KTH) database. The analysis demonstrates that the network training loss function reaches a minimum value of 0.0001. Leveraging the trained CNN model, the recognition accuracy for human motion behavior achieves peak performance, registering an average accuracy of 94.41%. Notably, the recognition accuracy for static motion behavior generally exceeds that for dynamic motion behavior across different models. The CNN-based human motion behavior recognition method exhibits promising results in both static and dynamic behavior recognition scenarios. Furthermore, the paper advocates for the use of IoT in collecting human motion behavior data in real-world living environments, contributing to the advancement of human motion behavior recognition technology and its application in diverse domains such as intelligent surveillance and health management. The research findings carry significant implications for furthering the development of human motion behavior recognition technology and enhancing its applications in areas such as intelligent surveillance and health management. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-08-07T15:15:54Z No. of bitstreams: 1 The Human Motion Behavior Recognition.pdf: 4269382 bytes, checksum: f53f0f7796ab666535653b73a3c92f2a (MD5); Made available in DSpace on 2024-08-07T15:15:54Z (GMT). No. of bitstreams: 1 The Human Motion Behavior Recognition.pdf: 4269382 bytes, checksum: f53f0f7796ab666535653b73a3c92f2a (MD5)

Liu, Jing-Wei https://reunir.unir.net/handle/123456789/17174 2024-09-03T11:49:36Z

Enhancing Tennis Serve Scoring Efficiency: An AI Deep Learning Approach Liu, Jing-Wei The playing field of a tennis competition is a dynamic and complex formative environment given the following preliminary knowledge: (a) the basic technical, tactical, situational, and special types of shots used by the opponent; (b) the hitting area of the tennis player; (c) the place of service; (d) the ball drop position; and (d) batting efficiency and other related information that may improve the chances of victory. In this study, we propose an AI classification model for tennis serve scores. Using a deep learning algorithm, the model automatically tracks and classifies the serve scores of professional tennis players from video data. We first defined the players’ techniques, volleys, and placements of strokes and serves. Subsequently, we defined the referee's tennis terms and the voice in deciding on a serve score. Finally, we developed a deep learning model to automatically classify the serving position, landing position, and use of tennis techniques. The methodology was applied in the context of 10 matches played by Roger Federer and Rafael Nadal. The proposed deep learning algorithm achieved a 98.27% accuracy in the automatic classification of serve scores, revealing that Nadal outscored Federer by 2.1% in terms of serve-scoring efficiency. These results are expected to facilitate the automatic comparison and classification of shots in future studies, enabling coaches to adjust tactics in a timely manner and thereby improve the chances of winning. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-08-07T09:57:58Z No. of bitstreams: 1 Enhancing Tennis Serve Scoring Efficiency.pdf: 605618 bytes, checksum: 3d89479d4e4a17f81ca2de5e1e9469e8 (MD5); Made available in DSpace on 2024-08-07T09:57:58Z (GMT). No. of bitstreams: 1 Enhancing Tennis Serve Scoring Efficiency.pdf: 605618 bytes, checksum: 3d89479d4e4a17f81ca2de5e1e9469e8 (MD5)

Montenegro-Marin, Carlos Enrique https://reunir.unir.net/handle/123456789/16747 2024-06-14T11:07:13Z

Editor’s Note Montenegro-Marin, Carlos Enrique This regular issue consists of 16 articles that use artificial intelligence or computational systems to come up with new solutions and solve problems more effectively. The issue showcases the use of Artificial Intelligence or computational systems that contribute to new knowledge with innovative applications. In this issue you can find different articles on game theory, models for collaborative filtering, text classification, fake news detection system, identification system, semi eager classifier, longitudinal segmented analysis, etc. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-06-14T11:07:13Z No. of bitstreams: 1 Editor’s Note.pdf: 56673 bytes, checksum: 6eb61f727eb6292a27ecad841527463a (MD5); Made available in DSpace on 2024-06-14T11:07:13Z (GMT). No. of bitstreams: 1 Editor’s Note.pdf: 56673 bytes, checksum: 6eb61f727eb6292a27ecad841527463a (MD5)

Dhahbi, Sami Saleem, Nasir Gunawan, Teddy Surya Bourouis, Sami Ali, Imad Trigui, Aymen Algarni, Abeer D. https://reunir.unir.net/handle/123456789/16570 2024-06-14T11:00:52Z

Lightweight Real-Time Recurrent Models for Speech Enhancement and Automatic Speech Recognition Dhahbi, Sami; Saleem, Nasir; Gunawan, Teddy Surya; Bourouis, Sami; Ali, Imad; Trigui, Aymen; Algarni, Abeer D. Traditional recurrent neural networks (RNNs) encounter difficulty in capturing long-term temporal dependencies. However, lightweight recurrent models for speech enhancement are important to improve noisy speech, while being computationally efficient and able to capture long-term temporal dependencies efficiently. This study proposes a lightweight hourglass-shaped model for speech enhancement (SE) and automatic speech recognition (ASR). Simple recurrent units (SRU) with skip connections are implemented where attention gates are added to the skip connections, highlighting the important features and spectral regions. The model operates without relying on future information that is well-suited for real-time processing. Combined acoustic features and two training objectives are estimated. Experimental evaluations using the short time speech intelligibility (STOI), perceptual evaluation of speech quality (PESQ), and word error rates (WERs) indicate better intelligibility, perceptual quality, and word recognition rates. The composite measures further confirm the performance of residual noise and speech distortion. With the TIMIT database, the proposed model improves the STOI and PESQ by 16.21% and 0.69 (31.1%) whereas with the LibriSpeech database, the model improves STOI by 16.41% and PESQ by 0.71 (32.9%) over the noisy speech. Further, our model outperforms other deep neural networks (DNNs) in seen and unseen conditions. The ASR performance is measured using the Kaldi toolkit and achieves 15.13% WERs in noisy backgrounds. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-05-13T16:16:38Z No. of bitstreams: 1 Lightweight Real-Time Recurrent Models for Speech Enhancement and Automatic Speech Recognition.pdf: 2470761 bytes, checksum: 74a0aaefed5f55e8bfcb6ef7c937544c (MD5); Made available in DSpace on 2024-05-13T16:16:38Z (GMT). No. of bitstreams: 1 Lightweight Real-Time Recurrent Models for Speech Enhancement and Automatic Speech Recognition.pdf: 2470761 bytes, checksum: 74a0aaefed5f55e8bfcb6ef7c937544c (MD5)

Su, Yu-Sheng Hu, Yu-Cheng Wu, Yun-Chin Lo, Ching-Teng https://reunir.unir.net/handle/123456789/16569 2024-09-03T11:54:46Z

Evaluating the Impact of Pumping on Groundwater Level Prediction in the Chuoshui River Alluvial Fan Using Artificial Intelligence Techniques Su, Yu-Sheng; Hu, Yu-Cheng; Wu, Yun-Chin; Lo, Ching-Teng Over the past decade, excessive groundwater extraction has been the leading cause of land subsidence in Taiwan's Chuoshui River Alluvial Fan (CRAF) area. To effectively manage and monitor groundwater resources, assessing the effects of varying seasonal groundwater extraction on groundwater levels is necessary. This study focuses on the CRAF in Taiwan. We applied three artificial intelligence techniques for three predictive models: multiple linear regression (MLR), support vector regression (SVR), and Long Short-Term Memory Networks (LSTM). Each prediction model evaluated the extraction rate, considering temporal and spatial correlations. The study aimed to predict groundwater level variations by comparing the results of different models. This study used groundwater level and extraction data from the CRAF area in Taiwan. The dataset we constructed was the input variable for predicting groundwater level variations. The experimental results show that the LSTM method is the most suitable and stable deep learning model for predicting groundwater level variations in the CRAF, Taiwan, followed by the SVR method and finally the MLR method. Additionally, when considering different distances and depths of pumping data at groundwater level monitoring stations, it was found that the Guosheng and Hexing groundwater level monitoring stations are best predicted using pumping data within a distance of 20 kilometers and a depth of 20 meters. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-05-13T16:01:02Z No. of bitstreams: 1 Evaluating the Impact of Pumping on Groundwater Level Prediction in the Chuoshui River Alluvial Fan Using Artificial Intelligence Techniques.pdf: 2739489 bytes, checksum: 8b91534008eeb41e6c20dd05d89810a1 (MD5); Made available in DSpace on 2024-05-13T16:01:02Z (GMT). No. of bitstreams: 1 Evaluating the Impact of Pumping on Groundwater Level Prediction in the Chuoshui River Alluvial Fan Using Artificial Intelligence Techniques.pdf: 2739489 bytes, checksum: 8b91534008eeb41e6c20dd05d89810a1 (MD5)

Chen, Chia-Chen Lin, Chia-Pei https://reunir.unir.net/handle/123456789/16264 2024-09-03T12:02:28Z

What Drives IoT-Based Smart Pet Appliances Usage Intention? The Perspective of the Unified Theory of Acceptance and Use of Technology Model Chen, Chia-Chen; Lin, Chia-Pei The advancement of IOT (Internet of Things) has facilitated the development of smart pet appliances, and the market for these products has growing rapidly, this study seeks to identify key factors for pet owner adoption of “smart” pet appliances. The Unified Theory of Acceptance and Use of Technology (UTAUT) a wellestablished model in the field of IOT research is used as the main framework, integrating brand trust, perceived value and perceived enjoyment as the basis for hypothesis formulation and testing based on data collected through questionnaires distributed through online social platforms. Reliability analysis, validity analysis and structural equation model analysis were carried out through confirmatory factor analysis to test the variables and research hypotheses. Results for the UTAUT indicate that effort expectancy has a direct impact on performance expectancy, while performance expectancy, effort expectancy and facilitating condition all have a positive impact on intention. While social influence does not directly or significantly affect use intention, it can indirectly affect intention through perceived value and perceived enjoyment. Brand trust does not have a significant impact on use intention, but can indirectly affect use intention through perceived value. This study further compares user age and number of smart pet home appliances owned to better understand the impact of demographic factors. Findings indicate that, for users under the age of 30, effort expectancy has no significant impact on use intention, while brand trust has no significant impact on perceived value among users over 30. Among the research results based on age as a basis, the impact of hardships in the ethnic group in the age of 30 is not significant, nor do facilitating conditions or perceived value have significant impact on use intention. For users with one smart pet device at home, neither favorable conditions not perceived value have significant impact on use intention, while for users with two smart pet devices, perceived enjoyment does not significantly impact use intention. These finding have potential reference value for future related research in the IOT or smart pet home appliance research field. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-04-04T09:47:52Z No. of bitstreams: 1 What Drives IoT-Based Smart Pet Appliances Usage Intention.pdf: 643943 bytes, checksum: efc27e1ab5afdf7df22a3f0a26d725b9 (MD5); Made available in DSpace on 2024-04-04T09:47:52Z (GMT). No. of bitstreams: 1 What Drives IoT-Based Smart Pet Appliances Usage Intention.pdf: 643943 bytes, checksum: efc27e1ab5afdf7df22a3f0a26d725b9 (MD5)

Alier, Marc García-Peñalvo, Francisco Camba, Jorge D. https://reunir.unir.net/handle/123456789/16211 2024-03-12T13:44:15Z

Generative Artificial Intelligence in Education: From Deceptive to Disruptive Alier, Marc; García-Peñalvo, Francisco; Camba, Jorge D. Generative Artificial Intelligence (GenAI) has emerged as a promising technology that can create original content, such as text, images, and sound. The use of GenAI in educational settings is becoming increasingly popular and offers a range of opportunities and challenges. This special issue explores the management and integration of GenAI in educational settings, including the ethical considerations, best practices, and opportunities. The potential of GenAI in education is vast. By using algorithms and data, GenAI can create original content that can be used to augment traditional teaching methods, creating a more interactive and personalized learning experience. In addition, GenAI can be utilized as an assessment tool and for providing feedback to students using generated content. For instance, it can be used to create custom quizzes, generate essay prompts, or even grade essays. The use of GenAI as an assessment tool can reduce the workload of teachers and help students receive prompt feedback on their work. Incorporating GenAI in educational settings also poses challenges related to academic integrity. With availability of GenAI models, students can use them to study or complete their homework assignments, which can raise concerns about the authenticity and authorship of the delivered work. Therefore, it is important to ensure that academic standards are maintained, and the originality of the student's work is preserved. This issue highlights the need for implementing ethical practices in the use of GenAI models and ensuring that the technology is used to support and not replace the student's learning experience. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T13:44:15Z No. of bitstreams: 1 Generative Artificial Intelligence in Education.pdf: 978561 bytes, checksum: 533e391d47ba81cd9a858d8b510f476a (MD5); Made available in DSpace on 2024-03-12T13:44:15Z (GMT). No. of bitstreams: 1 Generative Artificial Intelligence in Education.pdf: 978561 bytes, checksum: 533e391d47ba81cd9a858d8b510f476a (MD5)

Tang, Lin Su, Yu-Sheng https://reunir.unir.net/handle/123456789/16210 2024-03-12T13:30:39Z

Ethical Implications and Principles of Using Artificial Intelligence Models in the Classroom: A Systematic Literature Review Tang, Lin; Su, Yu-Sheng The increasing use of artificial intelligence (AI) models in the classroom not only brings a large number of benefits, but also has a variety of ethical implications. To provide effective education, it is now necessary to understand the ethical implications of using AI models in the classroom, and the principles for avoiding and addressing these ethical implications. However, existing research on the ethical implications of using AI models in the classroom is rather sparse, and a holistic overview is lacking. Therefore, this study seeks to offer an overview of research on the ethical implications, ethical principles and the future research directions and practices of using AI models in the classroom through a systematic literature review. Out of 1,445 initially identified publications between 2013 and 2023, 32 articles were included for final coding analysis, identified using explicit inclusion and exclusion criteria. The findings revealed five main ethical implications, namely algorithmic bias and discrimination, data privacy leakage, lack of transparency, decreased autonomy, and academic misconduct, with algorithmic bias being the most prominent (i.e., the number of existing studies is the most), followed by privacy leakage, whereas decreased autonomy and academic misconduct were relatively understudied; and six main ethical principles, namely fairness, privacy, transparency, accountability, autonomy and beneficence, with fairness being the most prominent ethical principle (i.e., the number of existing studies is the most), followed by privacy, while autonomy and beneficence were relatively understudied. Future directions of research are given, and guidelines for future practice are provided: (1) further substantive discussion, understanding and solution of ethical implications are required; (2) the precise mechanism of ethical principles of using AI models in the classroom remains to be elucidated and extended to the implementation phase; and (3) the ethical implications of the use of AI models in the classroom require accurate assessment. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T13:30:39Z No. of bitstreams: 1 Ethical Implications and Principles of Using Artificial Intelligence Models in the Classroom.pdf: 454767 bytes, checksum: 3722e4324633951952250b959168ccec (MD5); Made available in DSpace on 2024-03-12T13:30:39Z (GMT). No. of bitstreams: 1 Ethical Implications and Principles of Using Artificial Intelligence Models in the Classroom.pdf: 454767 bytes, checksum: 3722e4324633951952250b959168ccec (MD5)

Parra, Verónica Sureda, Patricia Corica, Ana Schiaffino, Silvia Godoy, Daniela https://reunir.unir.net/handle/123456789/16209 2024-03-12T13:21:58Z

Can Generative AI Solve Geometry Problems? Strengths and Weaknesses of LLMs for Geometric Reasoning in Spanish Parra, Verónica; Sureda, Patricia; Corica, Ana; Schiaffino, Silvia; Godoy, Daniela Generative Artificial Intelligence (AI) has emerged as a disruptive technology that is challenging traditional teaching and learning practices. Question-answering in natural language fosters the use of chatbots, such as ChatGPT, Bard and others, that generate text based on pre-trained Large Language Models (LLMs). The performance of these models in certain areas, like Math problem solving is receiving a crescent attention as it directly impacts on its potential use in educational settings. Most of these evaluations, however, concentrate on the construction and use of benchmarks comprising diverse Math problems in English. In this work, we discuss the capabilities of most used LLMs within the subfield of Geometry, in view of the relevance of this subject in high-school curricula and the difficulties exhibited by even most advanced multimodal LLMs to deal with geometric notions. This work focuses on Spanish, which is additionally a less resourced language. The answers of three major chatbots, based on different LLMs, were analyzed not only to determine their capacity to provide correct solutions, but also to categorize the errors found in the reasoning processes described. Understanding LLMs strengths and weaknesses in a field like Geometry can be a first step towards the design of more informed methodological proposals to include these technologies in classrooms as well as the development of more powerful automatic assistance tools based on generative AI. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T13:21:58Z No. of bitstreams: 1 Can Generative AI Solve Geometry Problems.pdf: 2743517 bytes, checksum: fe171389fc630d3009e3563d71f1ece1 (MD5); Made available in DSpace on 2024-03-12T13:21:58Z (GMT). No. of bitstreams: 1 Can Generative AI Solve Geometry Problems.pdf: 2743517 bytes, checksum: fe171389fc630d3009e3563d71f1ece1 (MD5)

Griffiths, Dai Frías-Martínez, Enrique Tlili, Ahmed Burgos, Daniel https://reunir.unir.net/handle/123456789/16207 2024-03-12T13:07:12Z

A Cybernetic Perspective on Generative AI in Education: From Transmission to Coordination Griffiths, Dai; Frías-Martínez, Enrique; Tlili, Ahmed; Burgos, Daniel The recent sudden increase in the capabilities of Large Language Models (LLMs), and generative AI in general, has astonished education professionals and learners. In formulating a response to these developments, educational institutions are constrained by a lack of clarity concerning human-machine communication and its relationship to models of education. Ideas and models from the cybernetic tradition can help to fill this gap. Two paradigms are distinguished: (1) the transmission paradigm (combining the model of learning implied by the instruments and processes of formal education and the conduit model of communication), and (2) the coordination paradigm (combining the constructivist model of learning and the coordination model of communication). It is proposed that these paradigms have long coexisted in educational practice in a modus vivendi, which is disrupted by LLMs. If an LLM can pass an examination, then from within the transmission paradigm this can only understood as demonstrating that the LLM has indeed learned and understood the material being assessed. At the same time, we know that LLMs do not in fact have the capacity to learn and understand, but rather generate a simulacrum of intelligence. It is argued that this paradox prevents educational institutions from formulating a coherent response to generative AI systems. However, within the coordination paradigm the interactions of LLMs and education institutions can be more easily understood and can be situated in a conversational model of learning. These distinctions can help institutions, educational leaders, and teachers, to frame the complex and nuanced questions raised by GenAI, and to chart a course towards its effective use in education. More specifically, they indicate that to benefit fully from the capabilities of generative AI education institutions need to recognize the validity of the coordination paradigm and adapt their processes and instruments accordingly. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T13:07:12Z No. of bitstreams: 1 A Cybernetic Perspective on Generative AI in Education.pdf: 547220 bytes, checksum: da8b47a5871354b7eada1c123edbb5cc (MD5); Made available in DSpace on 2024-03-12T13:07:12Z (GMT). No. of bitstreams: 1 A Cybernetic Perspective on Generative AI in Education.pdf: 547220 bytes, checksum: da8b47a5871354b7eada1c123edbb5cc (MD5)

Izquierdo-Domenech, Juan Linares-Pellicer, Jordi Ferri-Molla, Isabel https://reunir.unir.net/handle/123456789/16206 2024-03-12T12:54:12Z

Virtual Reality and Language Models, a New Frontier in Learning Izquierdo-Domenech, Juan; Linares-Pellicer, Jordi; Ferri-Molla, Isabel The proposed research introduces an innovative Virtual Reality (VR) and Large Language Model (LLM) architecture to enhance the learning process across diverse educational contexts, ranging from school to industrial settings. everaging the capabilities of LLMs and Retrieval-Augmented Generation (RAG), the architecture centers around an immersive VR application. This application empowers students of all backgrounds to interactively engage with their environment by posing questions and receiving informative responses in text format and with visual hints in VR, thereby fostering a dynamic learning experience. LLMs with RAG act as the backbones of this architecture, facilitating the integration of private or domain-specific data into the learning process. By seamlessly connecting various data sources through data connectors, RAG overcomes the challenge of disparate and siloed information repositories, including APIs, PDFs, SQL databases, and more. The data indexes provided by RAG solutions further streamline this process by structuring the ingested data into formats optimized for consumption by LLMs. An empirical study was conducted to evaluate the effectiveness of this VR and LLM architecture. Twenty participants, divided into Experimental and Control groups, were selected to assess the impact on their learning process. The Experimental group utilized the immersive VR application, which allowed interactive engagement with the educational environment, while the Control group followed traditional learning methods. The study revealed significant improvements in learning outcomes for the Experimental group, demonstrating the potential of integrating VR and LLMs in enhancing comprehension and engagement in learning contexts. This study presents an innovative approach that capitalizes on the synergy between LLMs and immersive VR technology, opening avenues for a transformative learning experience that transcends traditional boundaries and empowers learners across a spectrum of educational landscapes. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T12:54:12Z No. of bitstreams: 1 Virtual Reality and Language Models, a New Frontier in Learning.pdf: 2275154 bytes, checksum: 2dbd03feb08da813e2b458ff33488a46 (MD5); Made available in DSpace on 2024-03-12T12:54:12Z (GMT). No. of bitstreams: 1 Virtual Reality and Language Models, a New Frontier in Learning.pdf: 2275154 bytes, checksum: 2dbd03feb08da813e2b458ff33488a46 (MD5)

Bartlett, Kristin A. Camba, Jorge D. https://reunir.unir.net/handle/123456789/16205 2024-03-12T12:42:39Z

Generative Artificial Intelligence in Product Design Education: Navigating Concerns of Originality and Ethics Bartlett, Kristin A.; Camba, Jorge D. Image-generative artificial intelligence (AI) is increasingly being used in the product design process. In this paper, we present examples of how it is being used and discuss the possibilities of how applications may evolve in the future. We discuss the legal and ethical implications of image-generative AI, including concerns about bias, hidden labor, theft from artists, lack of originality in the outputs, and lack of copyright protection. We discuss how these concerns apply to design education and provide recommendations to educators about how AI should be addressed in the design classroom. We recommend that educators introduce AI as one tool among many in the designer’s toolkit and encourage it to be used as a process tool rather than for generating final design deliverables. We also provide guidance for how educators might engage students in discussions about AI to enhance their learning. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T12:42:39Z No. of bitstreams: 1 Generative Artificial Intelligence in Product Design Education.pdf: 2541510 bytes, checksum: 5ad68dd261b50c7854f81bbb23489f26 (MD5); Made available in DSpace on 2024-03-12T12:42:39Z (GMT). No. of bitstreams: 1 Generative Artificial Intelligence in Product Design Education.pdf: 2541510 bytes, checksum: 5ad68dd261b50c7854f81bbb23489f26 (MD5)

Rigaud Téllez, Nelly Rayón Villela, Patricia Blanco Bautista, Roberto https://reunir.unir.net/handle/123456789/16204 2024-03-12T12:18:25Z

Evaluating ChatGPT-Generated Linear Algebra Formative Assessments Rigaud Téllez, Nelly; Rayón Villela, Patricia; Blanco Bautista, Roberto This research explored Large Language Models potential uses on formative assessment for mathematical problem-solving process. The study provides a conceptual analysis of feedback and how the use of these models is related in the context of formative assessment for Linear Algebra problems. Particularly, the performance of a popular model known as ChatGPT in mathematical problems fails on reasoning, proofs, model construction, among others. Formative assessment is a process used by teachers and students during instruction that provides feedback to adjust ongoing teaching and learning to improve student’s achievement of intended instructional outcomes. The study analyzed and evaluated feedback provided to engineering students in their solutions, from both, instructors and ChatGPT, against fine-grained criteria of a formative feedback model that includes affective aspects. Considering preliminary outputs, and to improve performance of feedback from both agents’ instructors and ChatGPT, we developed a framework for formative assessment in mathematical problemsolving using a Large Language Model (LLM). We designed a framework to generate prompts, supported by common Linear Algebra mistakes within the context of concept development and problem-solving strategies. In this framework, the instructor acts as an agent to verify tasks in a math problem assigned to students, establishing a virtuous cycle of learning of queries supported by ChatGPT. Results revealed potentialities and challenges on how to improve feedback on graduate-level math problems, by which both educators and students adapt teaching and learning strategies. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T12:18:25Z No. of bitstreams: 1 Evaluating ChatGPT-Generated Linear Algebra Formative Assessments.pdf: 571024 bytes, checksum: c90c41bdd2f423882f07827c9f36cc78 (MD5); Made available in DSpace on 2024-03-12T12:18:25Z (GMT). No. of bitstreams: 1 Evaluating ChatGPT-Generated Linear Algebra Formative Assessments.pdf: 571024 bytes, checksum: c90c41bdd2f423882f07827c9f36cc78 (MD5)

Maslim, Martinus Wang, Hei-Chia Putra, Cendra Devayana Prabowo, Yulius Denny https://reunir.unir.net/handle/123456789/16203 2024-03-12T12:07:23Z

A Trustworthy Automated Short-Answer Scoring System Using a New Dataset and Hybrid Transfer Learning Method Maslim, Martinus; Wang, Hei-Chia; Putra, Cendra Devayana; Prabowo, Yulius Denny To measure the quality of student learning, teachers must conduct evaluations. One of the most efficient modes of evaluation is the short answer question. However, there can be inconsistencies in teacher-performed manual evaluations due to an excessive number of students, time demands, fatigue, etc. Consequently, teachers require a trustworthy system capable of autonomously and accurately evaluating student answers. Using hybrid transfer learning and student answer dataset, we aim to create a reliable automated short answer scoring system called Hybrid Transfer Learning for Automated Short Answer Scoring (HTL-ASAS). HTL-ASAS combines multiple tokenizers from a pretrained model with the bidirectional encoder representations from transformers. Based on our evaluation of the training model, we determined that HTL-ASAS has a higher evaluation accuracy than models used in previous studies. The accuracy of HTL-ASAS for datasets containing responses to questions pertaining to introductory information technology courses reaches 99.6%. With an accuracy close to one hundred percent, the developed model can undoubtedly serve as the foundation for a trustworthy ASAS system. Submitted by Andrea Sala (andrea.sala@unir.net) on 2024-03-12T09:23:17Z No. of bitstreams: 1 A Trustworthy Automated Short-Answer Scoring System Using a New Dataset and Hybrid Transfer Learning Method.pdf: 1131436 bytes, checksum: ad08b7933bf4c2c45c3cb3530a9f79c6 (MD5); Made available in DSpace on 2024-03-12T09:23:17Z (GMT). No. of bitstreams: 1 A Trustworthy Automated Short-Answer Scoring System Using a New Dataset and Hybrid Transfer Learning Method.pdf: 1131436 bytes, checksum: ad08b7933bf4c2c45c3cb3530a9f79c6 (MD5)

Hinderks, Andreas Domínguez Mayo, Francisco José Escalona, María José Thomaschewski, Jörg https://reunir.unir.net/handle/123456789/16005 2024-06-14T10:42:55Z

Requirements for User Experience Management - A Tertiary Study Hinderks, Andreas; Domínguez Mayo, Francisco José; Escalona, María José; Thomaschewski, Jörg Today’s users expect to be able to interact with the products they own without much effort and also want to be excited about them. The development of a positive user experience must therefore be managed. We understand management in general as a combination of a goal, a strategy, and resources. When applied to UX, user experience management consists of a UX goal, a UX strategy, and UX resources. We conducted a tertiary study and examined the current state of existing literature regarding possible requirements. We want to figure out, what requirements can be derived from the literature reviews with the focus on UX and agile development. In total, we were able to identify and analyse 16 studies. After analysing the studies in detail, we identified different requirements for UX management. In summary, we identified 13 requirements. The most frequently mentioned requirements were prototypes and UX/usability evaluation. Communication between UX professionals and developers was identified as a major improvement in the software development process. In summary, we were able to identify requirements for UX management of People/Social, Technology/Artifacts, and Process/Practice. However, we could not identify requirements for UX management that enabled the development and achievement of a UX goal. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2024-02-06T09:28:40Z No. of bitstreams: 1 ip2024_01_004.pdf: 707717 bytes, checksum: 636dc3688d95f8f658579c6594af7a52 (MD5); Made available in DSpace on 2024-02-06T09:28:40Z (GMT). No. of bitstreams: 1 ip2024_01_004.pdf: 707717 bytes, checksum: 636dc3688d95f8f658579c6594af7a52 (MD5)

Deore, Mahendra Tarambale, Manoj Raja Kumar, Jambi Ratna Sakhare, Sachin https://reunir.unir.net/handle/123456789/15785 2024-06-14T10:26:38Z

GRASE: Granulometry Analysis with Semi Eager Classifier to Detect Malware Deore, Mahendra; Tarambale, Manoj; Raja Kumar, Jambi Ratna; Sakhare, Sachin Technological advancement in communication leading to 5G, motivates everyone to get connected to the internet including ‘Devices’, a technology named Web of Things (WoT). The community benefits from this large-scale network which allows monitoring and controlling of physical devices. But many times, it costs the security as MALicious softWARE (MalWare) developers try to invade the network, as for them, these devices are like a ‘backdoor’ providing them easy ‘entry’. To stop invaders from entering the network, identifying malware and its variants is of great significance for cyberspace. Traditional methods of malware detection like static and dynamic ones, detect the malware but lack against new techniques used by malware developers like obfuscation, polymorphism and encryption. A machine learning approach to detect malware, where the classifier is trained with handcrafted features, is not potent against these techniques and asks for efforts to put in for the feature engineering. The paper proposes a malware classification using a visualization methodology wherein the disassembled malware code is transformed into grey images. It presents the efficacy of Granulometry texture analysis technique for improving malware classification. Furthermore, a Semi Eager (SemiE) classifier, which is a combination of eager learning and lazy learning technique, is used to get robust classification of malware families. The outcome of the experiment is promising since the proposed technique requires less training time to learn the semantics of higher-level malicious behaviours. Identifying the malware (testing phase) is also done faster. A benchmark database like malimg and Microsoft Malware Classification challenge (BIG-2015) has been utilized to analyse the performance of the system. An overall average classification accuracy of 99.03 and 99.11% is achieved, respectively. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2024-01-02T08:33:42Z No. of bitstreams: 1 ip2023_12_002_0.pdf: 3980674 bytes, checksum: 452dad3a10f3c7ee92f7982332310e05 (MD5); Made available in DSpace on 2024-01-02T08:33:42Z (GMT). No. of bitstreams: 1 ip2023_12_002_0.pdf: 3980674 bytes, checksum: 452dad3a10f3c7ee92f7982332310e05 (MD5)

Pérez-Antón, Raquel López-Sánchez, José Ignacio Corbi, Alberto https://reunir.unir.net/handle/123456789/15339 2024-06-14T10:19:03Z

The Game Theory in Quantum Computers: A Review Pérez-Antón, Raquel; López-Sánchez, José Ignacio; Corbi, Alberto Game theory has been studied extensively in recent centuries as a set of formal mathematical strategies for optimal decision making. This discipline improved its efficiency with the arrival, in the 20th century, of digital computer science. However, the computational limitations related to exponential time type problems in digital processors, triggered the search for more efficient alternatives. One of these choices is quantum computing. Certainly, quantum processors seem to be able to solve some of these complex problems, at least in theory. For this reason, in recent times, many research works have emerged related to the field of quantum game theory. In this paper we review the main studies about the subject, including operational requirements and implementation details. In addition, we describe various quantum games, their design strategy, and the used supporting tools. We also present the still open debate linked to the interpretation of the transformations of classical algorithms in fundamental game theory to their quantum version, with special attention to the Nash equilibrium. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-10-02T15:44:05Z No. of bitstreams: 1 ip2023_09_001.pdf: 1472228 bytes, checksum: 40c31750ef4e8ddfa372dabcd1fe17a8 (MD5); Made available in DSpace on 2023-10-02T15:44:05Z (GMT). No. of bitstreams: 1 ip2023_09_001.pdf: 1472228 bytes, checksum: 40c31750ef4e8ddfa372dabcd1fe17a8 (MD5)

Iqbal, Rashed Mokhlis, Hazlie Mohd Khairuddin, Anis Salwa Azam Muhammad, Munir https://reunir.unir.net/handle/123456789/15030 2025-04-24T14:53:14Z

An Improved Deep Learning Model for Electricity Price Forecasting Iqbal, Rashed; Mokhlis, Hazlie; Mohd Khairuddin, Anis Salwa; Azam Muhammad, Munir Accurate electricity price forecasting (EPF) is important for the purpose of bidding strategies and minimizing the risk for market participants in the competitive electricity market. Besides that, EPF becomes critically important for effective planning and efficient operation of a power system due to deregulation of electricity industry. However, accurate EPF is very challenging due to complex nonlinearity in the time series-based electricity prices. Hence, this work proposed two-fold contributions which are (1) effective time series preprocessing module to ensure feasible time-series data is fitted in the deep learning model, and (2) an improved long short-term memory (LSTM) model by incorporating linear scaled hyperbolic tangent (LiSHT) layer in the EPF. In this work, the time series pre-processing module adopted linear trend of the correlated features of electricity price series and the time series are tested by using Augmented Dickey Fuller (ADF) test method. In addition, the time series are transformed using boxcox transformation method in order to satisfy the stationarity property. Then, an improved LSTM prediction module is proposed to forecast electricity prices where LiSHT layer is adopted to optimize the parameters of the heterogeneous LSTM. This study is performed using the Australian electricity market price, load and renewable energy supply data. The experimental results obtained show that the proposed EPF framework performed better compared to previous techniques. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-07-11T10:59:08Z No. of bitstreams: 1 ip2023_06_001.pdf: 2496568 bytes, checksum: abb16105141372c6a491ec064dbf9a26 (MD5); Made available in DSpace on 2023-07-11T10:59:08Z (GMT). No. of bitstreams: 1 ip2023_06_001.pdf: 2496568 bytes, checksum: abb16105141372c6a491ec064dbf9a26 (MD5)

Fazal-E -Wahab Ye, Zhongfu Saleem, Nasir Ali, Hamza https://reunir.unir.net/handle/123456789/14813 2025-04-24T14:50:28Z

Efficient Gated Convolutional Recurrent Neural Networks for Real-Time Speech Enhancement Fazal-E -Wahab; Ye, Zhongfu; Saleem, Nasir; Ali, Hamza Deep learning (DL) networks have grown into powerful alternatives for speech enhancement and have achieved excellent results by improving speech quality, intelligibility, and background noise suppression. Due to high computational load, most of the DL models for speech enhancement are difficult to implement for realtime processing. It is challenging to formulate resource efficient and compact networks. In order to address this problem, we propose a resource efficient convolutional recurrent network to learn the complex ratio mask for real-time speech enhancement. Convolutional encoder-decoder and gated recurrent units (GRUs) are integrated into the Convolutional recurrent network architecture, thereby formulating a causal system appropriate for real-time speech processing. Parallel GRU grouping and efficient skipped connection techniques are engaged to achieve a compact network. In the proposed network, the causal encoder-decoder is composed of five convolutional (Conv2D) and deconvolutional (Deconv2D) layers. Leaky linear rectified unit (ReLU) is applied to all layers apart from the output layer where softplus activation to confine the network output to positive is utilized. Furthermore, batch normalization is adopted after every convolution (or deconvolution) and prior to activation. In the proposed network, different noise types and speakers can be used in training and testing. With the LibriSpeech dataset, the experiments show that the proposed real-time approach leads to improved objective perceptual quality and intelligibility with much fewer trainable parameters than existing LSTM and GRU models. The proposed model obtained an average of 83.53% STOI scores and 2.52 PESQ scores, respectively. The quality and intelligibility are improved by 31.61% and 17.18% respectively over noisy speech. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-06-01T10:22:12Z No. of bitstreams: 1 ip2023_05_007.pdf: 2253358 bytes, checksum: 2f2676659a89cc17ef33fee27b97df02 (MD5); Made available in DSpace on 2023-06-01T10:22:12Z (GMT). No. of bitstreams: 1 ip2023_05_007.pdf: 2253358 bytes, checksum: 2f2676659a89cc17ef33fee27b97df02 (MD5)

Meiners, Anna-Lena Schrepp, Martin Hinderks, Andreas Thomaschewski, Jörg https://reunir.unir.net/handle/123456789/14811 2025-04-24T14:44:46Z

A Benchmark for the UEQ+ Framework: Construction of a Simple Tool to Quickly Interpret UEQ+ KPIs Meiners, Anna-Lena; Schrepp, Martin; Hinderks, Andreas; Thomaschewski, Jörg Questionnaires are a highly efficient method to compare the user experience (UX) of different interactive products or versions of a single product. Concretely, they allow us to evaluate the UX easily and to compare different products with a numeric UX score. However, often only one UX score from a single evaluated product is available. Without a comparison to other measurements, it is difficult to interpret an individual score, e.g. to decide whether a product’s UX is good enough to compete in the market. Many questionnaires offer benchmarks to support researchers in these cases. A benchmark is the result of a larger set of product evaluations performed with the same questionnaire. The score obtained from a single product evaluation can be compared to the scores from this benchmark data set to quickly interpret the results. In this paper, the first benchmark for the UEQ+ (User Experience Questionnaire +) is presented, which was created using 3.290 UEQ+ responses for 26 successful software products. The UEQ+ is a modular framework that contains a high number of validated user experience scales that can be combined to form a UX questionnaire. Currently, no benchmark is available for this framework, making the benchmark constructed in this paper a valuable interpretation tool for UEQ+ questionnaires. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-06-01T10:03:31Z No. of bitstreams: 1 ip2023_05_003.pdf: 956451 bytes, checksum: 1ebeb9b7b4361cfbbd2175e1e172f038 (MD5); Made available in DSpace on 2023-06-01T10:03:31Z (GMT). No. of bitstreams: 1 ip2023_05_003.pdf: 956451 bytes, checksum: 1ebeb9b7b4361cfbbd2175e1e172f038 (MD5)

Caballero-Hernández, Juan Antonio Palomo-Duarte, Manuel Dodero, Juan Manuel Gaševic, Dragan https://reunir.unir.net/handle/123456789/14810 2024-06-14T10:12:09Z

Supporting Skill Assessment in Learning Experiences Based on Serious Games Through Process Mining Techniques Caballero-Hernández, Juan Antonio; Palomo-Duarte, Manuel; Dodero, Juan Manuel; Gaševic, Dragan Learning experiences based on serious games are employed in multiple contexts. Players carry out multiple interactions during the gameplay to solve the different challenges faced. Those interactions can be registered in logs as large data sets providing the assessment process with objective information about the skills employed. Most assessment methods in learning experiences based on serious games rely on manual approaches, which do not scalewell when the amount of data increases. We propose an automated method to analyse students’ interactions and assess their skills in learning experiences based on serious games. The method takes into account not only the final model obtained by the student, but also the process followed to obtain it, extracted from game logs. The assessment method groups students according to their in-game errors and ingame outcomes. Then, the models for the most and the least successful students are discovered using process mining techniques. Similarities in their behaviour are analysed through conformance checking techniques to compare all the students with the most successful ones. Finally, the similarities found are quantified to build a classification of the students’ assessments. We have employed this method with Computer Science students playing a serious game to solve design problems in a course on databases. The findings show that process mining techniques can palliate the limitations of skill assessment methods in game-based learning experiences. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-06-01T09:53:03Z No. of bitstreams: 1 ip2023_05_002.pdf: 2378538 bytes, checksum: cbb82497cc2949dd439295c221f7095d (MD5); Made available in DSpace on 2023-06-01T09:53:03Z (GMT). No. of bitstreams: 1 ip2023_05_002.pdf: 2378538 bytes, checksum: cbb82497cc2949dd439295c221f7095d (MD5)

García García, Miguel González Arrieta, María Angélica Rodríguez González, Sara Márquez-Sánchez, Sergio Da Silva Ramos, Carlos Fernando https://reunir.unir.net/handle/123456789/14809 2024-06-14T09:54:40Z

Graffiti Identification System Using Low-Cost Sensors García García, Miguel; González Arrieta, María Angélica; Rodríguez González, Sara; Márquez-Sánchez, Sergio; Da Silva Ramos, Carlos Fernando This article introduces the possibility of studying graffiti using a colorimeter developed with Arduino hardware technology according to the Do It Yourself (DIY) philosophy. Through the obtained Red Green Blue (RGB) data it is intended to study and compare the information extracted from each of the graffiti present on different walls. The same color can be found in different parts of a single graffiti, but also in other graffiti that could a priori be of different authorship. Nevertheless, graffiti may be related, and it may be possible to group graffiti artists and "gangs" that work together. The methodology followed for the construction of the colorimeter and its real application in a practical case are described in four case studies. The case studies describe how graffiti were identified and recognized and they provide a comparison of the collected color samples. The results show the added value of the colorimeter in the graffiti recognition process, demonstrating its usefulness on a functional level. Finally, the contributions of this research are outlined, and an analysis is carried out of the changes to be made to the proposed method in the future, for improved graffiti color identification. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-06-01T09:42:02Z No. of bitstreams: 1 ip2023_05_001.pdf: 4129896 bytes, checksum: 1327ead89f0c7692e534ae14e61b9f4b (MD5); Made available in DSpace on 2023-06-01T09:42:02Z (GMT). No. of bitstreams: 1 ip2023_05_001.pdf: 4129896 bytes, checksum: 1327ead89f0c7692e534ae14e61b9f4b (MD5)

Bobadilla, Jesús Dueñas-Lerín, Jorge Ortega, Fernando Gutiérrez, Abraham https://reunir.unir.net/handle/123456789/14594 2024-06-14T09:47:10Z

Comprehensive Evaluation of Matrix Factorization Models for Collaborative Filtering Recommender Systems Bobadilla, Jesús; Dueñas-Lerín, Jorge; Ortega, Fernando; Gutiérrez, Abraham Matrix factorization models are the core of current commercial collaborative filtering Recommender Systems. This paper tested six representative matrix factorization models, using four collaborative filtering datasets. Experiments have tested a variety of accuracy and beyond accuracy quality measures, including prediction, recommendation of ordered and unordered lists, novelty, and diversity. Results show each convenient matrix factorization model attending to their simplicity, the required prediction quality, the necessary recommendation quality, the desired recommendation novelty and diversity, the need to explain recommendations, the adequacy of assigning semantic interpretations to hidden factors, the advisability of recommending to groups of users, and the need to obtain reliability values. To ensure the reproducibility of the experiments, an open framework has been used, and the implementation code is provided. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-05-03T10:32:26Z No. of bitstreams: 1 ip2023_04_008.pdf: 6298241 bytes, checksum: 78a2b02200b2d8058f71baecdc8062ba (MD5); Made available in DSpace on 2023-05-03T10:32:26Z (GMT). No. of bitstreams: 1 ip2023_04_008.pdf: 6298241 bytes, checksum: 78a2b02200b2d8058f71baecdc8062ba (MD5)

Manresa-Yee, Cristina Ramis, Silvia Buades, José M. https://reunir.unir.net/handle/123456789/14589 2025-04-24T14:46:13Z

Analysis of Gender Differences in Facial Expression Recognition Based on Deep Learning Using Explainable Artificial Intelligence Manresa-Yee, Cristina; Ramis, Silvia; Buades, José M. Potential uses of automated Facial Expression Recognition (FER) cover a wide range of applications such as customer behavior analysis, healthcare applications or providing personalized services. Data for machine learning play a fundamental role, therefore, understanding the relevancy of the data in the outcomes is of utmost importance. In this work we present a study on how gender influences the learning of a FER system. We analyze with Explainable Artificial intelligence (XAI) techniques how gender contributes to the learning and assess which facial expressions are more similar regarding face regions that impact on the classification. Results show that there exist common regions in some expressions both for females and males with different intensities (e.g. happiness); however, there are other expressions like disgust, where important face regions differ. The insights of this work will help improving FER systems and understand the source of any inequality. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-05-03T09:22:35Z No. of bitstreams: 1 ip2023_04_003.pdf: 2400476 bytes, checksum: 29a450353bf34f3a9da5ddb4e01f8a22 (MD5); Made available in DSpace on 2023-05-03T09:22:35Z (GMT). No. of bitstreams: 1 ip2023_04_003.pdf: 2400476 bytes, checksum: 29a450353bf34f3a9da5ddb4e01f8a22 (MD5)

Tomar, A. Kumar, S. Pant, B. https://reunir.unir.net/handle/123456789/14588 2024-06-14T09:40:17Z

PeopleNet: A Novel People Counting Framework for Head-Mounted Moving Camera Videos Tomar, A.; Kumar, S.; Pant, B. Traditional crowd counting (optical flow or feature matching) techniques have been upgraded to deep learning (DL) models due to their lack of automatic feature extraction and low-precision outcomes. Most of these models were tested on surveillance scene crowd datasets captured by stationary shooting equipment. It is very challenging to perform people counting from the videos shot with a head-mounted moving camera; this is mainly due to mixing the temporal information of the moving crowd with the induced camera motion. This study proposed a transfer learning-based PeopleNet model to tackle this significant problem. For this, we have made some significant changes to the standard VGG16 model, by disabling top convolutional blocks and replacing its standard fully connected layers with some new fully connected and dense layers. The strong transfer learning capability of the VGG16 network yields in-depth insights of the PeopleNet into the good quality of density maps resulting in highly accurate crowd estimation. The performance of the proposed model has been tested over a self-generated image database prepared from moving camera video clips, as there is no public and benchmark dataset for this work. The proposed framework has given promising results on various crowd categories such as dense, sparse, average, etc. To ensure versatility, we have done self and cross-evaluation on various crowd counting models and datasets, which proves the importance of the PeopleNet model in adverse defense of society. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-05-03T09:09:31Z No. of bitstreams: 1 ip2023_04_002_.pdf: 6031442 bytes, checksum: fd5da7b99c7eabd3a996764acecebd7f (MD5); Made available in DSpace on 2023-05-03T09:09:31Z (GMT). No. of bitstreams: 1 ip2023_04_002_.pdf: 6031442 bytes, checksum: fd5da7b99c7eabd3a996764acecebd7f (MD5)

Cervantes, Alejandro Quintana, David Saez, Yago Isasi, Pedro https://reunir.unir.net/handle/123456789/14369 2024-06-13T15:17:02Z

Longitudinal Segmented Analysis of Internet Usage and Well-Being Among Older Adults Cervantes, Alejandro; Quintana, David; Saez, Yago; Isasi, Pedro The connection between digital literacy and the three core dimensions of psychological well-being is not yet well understood, and the evidence is controversial. We analyzed a sample of 2,314 individuals, aged 50 years and older, that participated in the English Longitudinal Study of Aging. Participants were clustered according to drivers of psychological well-being using Self-Organizing Maps. The resulting groups were subsequently studied separately using generalized estimating equations fitted on 2-year lagged repeated measures using three scales to capture the dimensions of well-being and Markov models. The clustering analysis suggested the existence of four different groups of participants. Statistical models found differences in the connection between internet use and psychological well-being depending on the group. The Markov models showed a clear association between internet use and the potential for transition among groups of the population characterized, among other things, by higher levels of psychological well-being. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-15T14:19:08Z No. of bitstreams: 1 ip2022_05_002.pdf: 742179 bytes, checksum: 9e2c29ca131f0bcbd368e0c12f8e11cd (MD5); Made available in DSpace on 2023-03-15T14:19:08Z (GMT). No. of bitstreams: 1 ip2022_05_002.pdf: 742179 bytes, checksum: 9e2c29ca131f0bcbd368e0c12f8e11cd (MD5)

Ribelles, Jose Lopez, Angeles Traver, V. Javier https://reunir.unir.net/handle/123456789/14364 2024-06-13T15:18:23Z

Modulating the Gameplay Challenge Through Simple Visual Computing Elements: A Cube Puzzle Case Study Ribelles, Jose; Lopez, Angeles; Traver, V. Javier Positive player’s experiences greatly rely on a balanced gameplay where the game difficulty is related to player’s skill. Towards this goal, the gameplay can be modulated to make it easier or harder. In this work, a modulating mechanism based on visual computing is explored. The main hypothesis is that simple visual modifications of some elements in the game can have a significant impact on the game experience. This concept, which is essentially unexplored in the literature, has been experimentally tested with a web-based cube puzzle game where participants played either the original game or the visually modified game. The analysis is based on players’ behavior, performance, and replies to a questionnaire upon game completion. The results provide evidence on the effectiveness of visual computing on gameplay modulation. We believe the findings are relevant to game researchers and developers because they highlight how a core gameplay can be easily modified with relatively simple ingredients, at least for some game genres. Interestingly, the insights gained from this study also open the door to automate the game adaptation based on observed player’s interaction. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-15T09:14:27Z No. of bitstreams: 1 ip2022_05_001.pdf: 5360016 bytes, checksum: cad2037fa92328670eb0032ba3e802f1 (MD5); Made available in DSpace on 2023-03-15T09:14:27Z (GMT). No. of bitstreams: 1 ip2022_05_001.pdf: 5360016 bytes, checksum: cad2037fa92328670eb0032ba3e802f1 (MD5)

Suruliandi, A. Idhaya, T. Raja, S. P. https://reunir.unir.net/handle/123456789/14363 2024-06-13T15:51:43Z

Drug Target Interaction Prediction Using Machine Learning Techniques – A Review Suruliandi, A.; Idhaya, T.; Raja, S. P. Drug discovery is a key process, given the rising and ubiquitous demand for medication to stay in good shape right through the course of one’s life. Drugs are small molecules that inhibit or activate the function of a protein, offering patients a host of therapeutic benefits. Drug design is the inventive process of finding new medication, based on targets or proteins. Identifying new drugs is a process that involves time and money. This is where computer-aided drug design helps cut time and costs. Drug design needs drug targets that are a protein and a drug compound, with which the interaction between a drug and a target is established. Interaction, in this context, refers to the process of discovering protein binding sites, which are protein pockets that bind with drugs. Pockets are regions on a protein macromolecule that bind to drug molecules. Researchers have been at work trying to determine new Drug Target Interactions (DTI) that predict whether or not a given drug molecule will bind to a target. Machine learning (ML) techniques help establish the interaction between drugs and their targets, using computer-aided drug design. This paper aims to explore ML techniques better for DTI prediction and boost future research. Qualitative and quantitative analyses of ML techniques show that several have been applied to predict DTIs, employing a range of classifiers. Though DTI prediction improves with negative drug target pairs (DTP), the lack of true negative DTPs has led to the use a particular dataset of drugs and targets. Using dynamic DTPs improves DTI prediction. Little attention has so far been paid to developing a new classifier for DTI classification, and there is, unquestionably, a need for better ones. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-15T08:58:46Z No. of bitstreams: 1 ip2022_11_002_0.pdf: 920478 bytes, checksum: 045180cf6b9ca6c626f93afc48d23879 (MD5); Made available in DSpace on 2023-03-15T08:58:46Z (GMT). No. of bitstreams: 1 ip2022_11_002_0.pdf: 920478 bytes, checksum: 045180cf6b9ca6c626f93afc48d23879 (MD5)

del Águila Escobar, Raúl A. Suárez-Figueroa, Mari Carmen Fernández-López, Mariano https://reunir.unir.net/handle/123456789/14362 2024-06-13T15:44:39Z

OBOE: an Explainable Text Classification Framework del Águila Escobar, Raúl A.; Suárez-Figueroa, Mari Carmen; Fernández-López, Mariano Explainable Artificial Intelligence (XAI) has recently gained visibility as one of the main topics of Artificial Intelligence research due to, among others, the need to provide a meaningful justification of the reasons behind the decision of black-box algorithms. Current approaches are based on model agnostic or ad-hoc solutions and, although there are frameworks that define workflows to generate meaningful explanations, a text classification framework that provides such explanations considering the different ingredients involved in the classification process (data, model, explanations, and users) is still missing. With the intention of covering this research gap, in this paper we present a text classification framework called OBOE (explanatiOns Based On concEpts), in which such ingredients play an active role to open the black-box. OBOE defines different components whose implementation can be customized and, thus, explanations are adapted to specific contexts. We also provide a tailored implementation to show the customization capability of OBOE. Additionally, we performed (a) a validation of the implemented framework to evaluate the performance using different corpora and (b) a user-based evaluation of the explanations provided by OBOE. The latter evaluation shows that the explanations generated in natural language express the reason for the classification results in a way that is comprehensible to non-technical users. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-15T08:40:46Z No. of bitstreams: 1 ip2022_11_01.pdf: 1457451 bytes, checksum: 5fd3f5aeda8a86102d1df7f228ab1c4a (MD5); Made available in DSpace on 2023-03-15T08:40:46Z (GMT). No. of bitstreams: 1 ip2022_11_01.pdf: 1457451 bytes, checksum: 5fd3f5aeda8a86102d1df7f228ab1c4a (MD5)

Kumaar, M. Akshay Samiayya, Duraimurugan Rajinikanth, Venkatesan Raj Vincent P M, Durai Kadry, Seifedine https://reunir.unir.net/handle/123456789/14357 2024-06-14T09:32:03Z

Brain Tumor Classification Using a Pre-Trained Auxiliary Classifying Style-Based Generative Adversarial Network Kumaar, M. Akshay; Samiayya, Duraimurugan; Rajinikanth, Venkatesan; Raj Vincent P M, Durai; Kadry, Seifedine Computer Vision's applications and their use cases in the medical field have grown vastly in the past decade. The algorithms involved in these critical applications have helped doctors and surgeons perform procedures on patients more precisely with minimal side effects. However, obtaining medical data for developing large scale generalizable and intelligent algorithms is challenging in the real world as multiple socio-economic, administrative, and demographic factors impact it. Furthermore, training machine learning algorithms with a small amount of data can lead to less accuracy and performance bias, resulting in incorrect diagnosis and treatment, which can cause severe side effects or even casualties. Generative Adversarial Networks (GAN) have recently proven to be an effective data synthesis and augmentation technique for training deep learning-based image classifiers. This research proposes a novel approach that uses a Style-based Generative Adversarial Network for conditional synthesis and auxiliary classification of Brain Tumors by pre-training. The Discriminator of the pre-trained GAN is fine-tuned with extensive data augmentation techniques to improve the classification accuracy when the training data is small. The proposed method was validated with an open-source MRI dataset which consists of three types of tumors - Glioma, Meningioma, and Pituitary. The proposed system achieved 99.51% test accuracy, 99.52% precision score, and 99.50% recall score, significantly higher than other approaches. Since the framework can be made adaptive using transfer learning, this method also benefits new and small datasets of similar distributions. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-14T11:08:15Z No. of bitstreams: 1 ip2023_02_008_0.pdf: 2381727 bytes, checksum: 316091571fa76e03f42a6c6be5902ef2 (MD5); Made available in DSpace on 2023-03-14T11:08:15Z (GMT). No. of bitstreams: 1 ip2023_02_008_0.pdf: 2381727 bytes, checksum: 316091571fa76e03f42a6c6be5902ef2 (MD5)

Ali Reshi, Junaid Ali, Rashid https://reunir.unir.net/handle/123456789/14339 2024-06-14T09:25:29Z

An Efficient Fake News Detection System Using Contextualized Embeddings and Recurrent Neural Network Ali Reshi, Junaid; Ali, Rashid Fake news is detrimental for society and individuals. Since the information dissipation through online media is too quick, an efficient system is needed to detect and counter the propagation of fake news on social media. Many studies have been performed in last few years to detect fake news on social media. This study focusses on the efficient detection of fake news on social media, through a Natural Language Processing based approach, using deep learning. For the detection of fake news, textual data have been analyzed in unidirectional way using sequential neural networks, or in bi-directional way using transformer architectures like Bidirectional Encoder Representations from Transformers (BERT). This paper proposes ConFaDe - a deep learning based fake news detection system that utilizes contextual embeddings generated from a transformer-based model. The model uses Masked Language Modelling and Replaced Token Detection in its pre-training to capture contextual and semantic information in the text. The proposed system outperforms the previously set benchmarks for fake news detection; including state-of-the-art approaches on a real-world fake news dataset, when evaluated using a set of standard performance metrics with an accuracy of 99.9 % and F1 macro of 99.9%. In contrast to the existing state-of-the-art model, the proposed system uses 90 percent less network parameters and is 75 percent lesser in size. Consequently, ConFaDe requires fewer hardware resources and less training time, and yet outperforms the existing fake news detection techniques, a step forward in the direction of Green Artificial Intelligence. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-13T12:06:50Z No. of bitstreams: 1 ip2023_02_007_1.pdf: 1828984 bytes, checksum: 10970f3c192f73e3f2f4b1b0f7068095 (MD5); Made available in DSpace on 2023-03-13T12:06:50Z (GMT). No. of bitstreams: 1 ip2023_02_007_1.pdf: 1828984 bytes, checksum: 10970f3c192f73e3f2f4b1b0f7068095 (MD5)

García-Peñalvo, Francisco Vázquez-Ingelmo, Andrea García-Holgado, Alicia Sampedro-Gómez, Jesús Sánchez-Puente, Antonio Vicente-Palacios, Víctor Dorado-Díaz, P. Ignacio Sánchez, Pedro L. https://reunir.unir.net/handle/123456789/14314 2024-06-14T09:14:37Z

KoopaML: A Graphical Platform for Building Machine Learning Pipelines Adapted to Health Professionals García-Peñalvo, Francisco; Vázquez-Ingelmo, Andrea; García-Holgado, Alicia; Sampedro-Gómez, Jesús; Sánchez-Puente, Antonio; Vicente-Palacios, Víctor; Dorado-Díaz, P. Ignacio; Sánchez, Pedro L. Machine Learning (ML) has extended its use in several domains to support complex analyses of data. The medical field, in which significant quantities of data are continuously generated, is one of the domains that can benefit from the application of ML pipelines to solve specific problems such as diagnosis, classification, disease detection, segmentation, assessment of organ functions, etc. However, while health professionals are experts in their domain, they can lack programming and theoretical skills regarding ML applications. Therefore, it is necessary to train health professionals in using these paradigms to get the most out of the application of ML algorithms to their data. In this work, we present a platform to assist non-expert users in defining ML pipelines in the health domain. The system’s design focuses on providing an educational experience to understand how ML algorithms work and how to interpret their outcomes and on fostering a flexible architecture to allow the evolution of the available components, algorithms, and heuristics. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-09T15:55:31Z No. of bitstreams: 1 ip2023_01_006.pdf: 1776887 bytes, checksum: 6df5ba28ca606edf3547654db4c11ff0 (MD5); Made available in DSpace on 2023-03-09T15:55:31Z (GMT). No. of bitstreams: 1 ip2023_01_006.pdf: 1776887 bytes, checksum: 6df5ba28ca606edf3547654db4c11ff0 (MD5)

Khrissi, Lahbib El Akkad, Nabil Satori, Hassan Satori, Khalid https://reunir.unir.net/handle/123456789/14313 2025-04-24T14:47:29Z

A Feature Selection Approach Based on Archimedes’ Optimization Algorithm for Optimal Data Classification Khrissi, Lahbib; El Akkad, Nabil; Satori, Hassan; Satori, Khalid Feature selection is an active research area in data mining and machine learning, especially with the increase in the amount of numerical data. FS is a search strategy to find the best subset of features among a large number of subsets of features. Thus, FS is applied in most modern applications and in various domains, which requires the search for a powerful FS technique to process and classify high-dimensional data. In this paper, we propose a new technique for dimension reduction in feature selection. This approach is based on a recent metaheuristic called Archimedes’ Optimization Algorithm (AOA) to select an optimal subset of features to improve the classification accuracy. The idea of the AOA is based on the steps of Archimedes' principle in physics. It explains the behavior of the force exerted when an object is partially or fully immersed in a fluid. AOA optimization maintains a balance between exploration and exploitation, keeping a population of solutions and studying a large area to find the best overall solution. In this study, AOA is exploited as a search technique to find an optimal feature subset that reduces the number of features to maximize classification accuracy. The K-nearest neighbor (K-NN) classifier was used to evaluate the classification performance of selected feature subsets. To demonstrate the superiority of the proposed method, 16 benchmark datasets from the UCI repository are used and also compared by well-known and recently introduced meta-heuristics in this context, such as: sine-cosine algorithm (SCA), whale optimization algorithm (WOA), butterfly optimization algorithm (BAO), and butterfly flame optimization algorithm (MFO). The results prove the effectiveness of the proposed algorithm over the other algorithms based on several performance measures used in this paper. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-09T15:45:11Z No. of bitstreams: 1 ip2023_01_005.pdf: 1715840 bytes, checksum: db554ba969c6df88b4c141b7b1499932 (MD5); Made available in DSpace on 2023-03-09T15:45:11Z (GMT). No. of bitstreams: 1 ip2023_01_005.pdf: 1715840 bytes, checksum: db554ba969c6df88b4c141b7b1499932 (MD5)

Balderas, Antonio Baena-Pérez, Rubén Person, Tatiana Mota, José Miguel Ruiz-Rube, Iván https://reunir.unir.net/handle/123456789/14311 2024-06-13T15:34:19Z

Chatbot-Based Learning Platform for SQL Training Balderas, Antonio; Baena-Pérez, Rubén; Person, Tatiana; Mota, José Miguel; Ruiz-Rube, Iván Learning the SQL language for working with relational databases is a fundamental subject for future computer engineers. However, in distance learning contexts or unexpected situations like the COVID-19 pandemic, where students had to follow lectures remotely, they may find it hard to learn. Chatbots are software applications that aim to have conversations with people to help them solve problems or provide support in a specific domain. This paper proposes a chatbot-based learning platform to assist students in learning SQL. A case study has been conducted to evaluate the proposal, with undergraduate computer engineering students using the learning platform to perform SQL queries while being assisted by the chatbot. The results show evidence that students who used the chatbot performed better on the final SQL exam than those who did not. In addition, the research shows positive evidence of the benefits of using such learning platforms to support SQL teaching and learning for both students and lecturers: students use a platform that helps them self-regulate their learning process, while lecturers get interesting metrics on student performance. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2023-03-09T15:01:20Z No. of bitstreams: 1 ip2022_05_003.pdf: 1402528 bytes, checksum: fa631e1bc6e7d15b9f36bd1e2d12c703 (MD5); Made available in DSpace on 2023-03-09T15:01:20Z (GMT). No. of bitstreams: 1 ip2022_05_003.pdf: 1402528 bytes, checksum: fa631e1bc6e7d15b9f36bd1e2d12c703 (MD5)