https://reunir.unir.net/handle/123456789/13133 Thu, 26 Feb 2026 15:05:49 GMT 2026-02-26T15:05:49Z https://reunir.unir.net/handle/123456789/13151 Research Collaboration Influence Analysis Using Dynamic Co-authorship and Citation Networks Razzaq, Sidra; Kamran Malik, Ahmad; Raza, Basit; Ali Khattak, Hasan; Moscoso Zegarra, Giomar W.; Díaz Zelada, Yvan Collaborative research is increasing in terms of publications, skills, and formal interactions, which certainly makes it the hotspot in both academia and the industrial sector. Knowing the factors and behavior of dynamic collaboration network provides insights that helps in improving the researcher’s profile and coordinator’s productivity of research. Despite rapid developments in the research collaboration process with various outcomes, its validity is still difficult to address. Existing approaches have used bibliometric network analysis with different aspects to understand collaboration patterns that measure the quality of their corresponding relationships. At this point in time, we would like to investigate an efficient method to outline the credibility of findings in publication—author relations. In this research, we propose a new collaboration method to analyze the structure of research articles using four types of graphs for discerning authors’ influence. We apply different combinations of network relationships and bibliometric analysis on the G-index parameter to disclose their interrelated differences. Our model is designed to find the dynamic indicators of co-authored collaboration with an influence on the author’s behavior in terms of change in research area/interest. In the research we investigate the dynamic relations in an academic field using metadata of openly available articles and collaborating international authors in interrelated areas/domains. Based on filtered evidence of relationship networks and their statistical results, the research shows an increment in productivity and better influence over time. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-23T08:56:54Z No. of bitstreams: 1 ijimai7_3_10_0.pdf: 1917831 bytes, checksum: 18756ee4e984bc423bb9c79843b0ee00 (MD5); Made available in DSpace on 2022-05-23T08:56:54Z (GMT). No. of bitstreams: 1 ijimai7_3_10_0.pdf: 1917831 bytes, checksum: 18756ee4e984bc423bb9c79843b0ee00 (MD5) https://reunir.unir.net/handle/123456789/13151 https://reunir.unir.net/handle/123456789/13150 Machine Learning in Business Intelligence 4.0: Cost Control in a Destination Hotel Sánchez-Torres, Fulgencio; González, Iván; Dobrescu, Cosmin C. Cost control is a recurring problem in companies where studies have provided different solutions. The main objective of this research is to propose and validate an alternative to cost control using data science to support decision-making using the business intelligence 4.0 paradigm. The work uses Machine Learning (ML) to support decision-making in company cost-control management. Specifically, we used the ability of hierarchical agglomerative clustering (HAC) algorithms to generate clusters and suggest possible candidate products that could be substituted for other, more cost-effective ones. These candidate products were analyzed by a panel of company experts, facilitating decisions based on business costs. We needed to analyze and modify the company's ecosystem and its associated variables to obtain an adequate data warehouse during the study, which was developed over three years and validated HAC as a support to decision-making in cost control. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-23T08:50:20Z No. of bitstreams: 1 ijimai7_3_8_0.pdf: 662619 bytes, checksum: e14c1821a0007cbf28e6fe80ba05e3d4 (MD5); Made available in DSpace on 2022-05-23T08:50:20Z (GMT). No. of bitstreams: 1 ijimai7_3_8_0.pdf: 662619 bytes, checksum: e14c1821a0007cbf28e6fe80ba05e3d4 (MD5) https://reunir.unir.net/handle/123456789/13150 https://reunir.unir.net/handle/123456789/13149 Editor's Note Golpe, Antonio A.; Isasi, Pedro; Martín-Álvarez, Juan Manuel; Quintana, David Machine learning (ML) is generating new opportunities for innovative research in areas apparently unrelated such as, economics, business or/and finance. Specifically, it has also been widely used in applications related to the economic and financial analysis, such as economic recessions prediction, labor market trends, risk management, prices analysis among others. However, it is important to note the differences between classical statistics/econometrics and machine learning. On the one hand, econometrics set out to build models designed to describe economic problems, while machine learning uses algorithms, generally for prediction, classification, and also, can manage a large amount of structured and unstructured data and make fast decisions or forecasts. As S. Athey points out, perhaps “a key advantage of ML is that it frames empirical analysis in terms of algorithms that estimate and compare many alternative models. This approach contrasts with econometrics, where (in principle, though rarely in reality) the researcher picks a model based on principles and estimates it once”. This Special Issue presents nine contributions that illustrate both approaches in the domain of economics, finance and business. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-23T08:12:34Z No. of bitstreams: 1 ijimai7_3_0.pdf: 58542 bytes, checksum: 521ed0728cdb1a2bd1e04486981d07b4 (MD5); Made available in DSpace on 2022-05-23T08:12:34Z (GMT). No. of bitstreams: 1 ijimai7_3_0.pdf: 58542 bytes, checksum: 521ed0728cdb1a2bd1e04486981d07b4 (MD5) https://reunir.unir.net/handle/123456789/13149 https://reunir.unir.net/handle/123456789/13148 Foreword Velarde Molina, Jehovanni Fabricio This time, in the Special Issue on Artificial Intelligence in Economics, Finance and Business, we present a series of publications focused on artificial intelligence and finance. This compilation of research will bring new information to researchers in different disciplines, and at the same time, it will be an ideal space to present studies that have an international scope. UNIR, dedicated to the training of professionals in different academic programs, through its journal is consolidating a culture of research and expanding the knowledge that contributes to an excellent education. For this reason, we consider the dissemination of scientific articles essential, since this guarantees the transfer of results, in addition to the conclusions of high-impact research. Currently the world is going through a complicated scenario, a fluctuating economy and problems in health services that require immediate attention; in this sense, science and knowledge management open space to opportunities in search of medium and long-term solutions. It is a great honor to present this issue of the International Journal of Interactive Multimedia and Artificial Intelligence, whose contribution to the knowledge society is invaluable. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-23T07:57:14Z No. of bitstreams: 1 ijimai7_3_00.pdf: 59365 bytes, checksum: 8d7e35f96ee9570c6b66e9776f9b0049 (MD5); Made available in DSpace on 2022-05-23T07:57:14Z (GMT). No. of bitstreams: 1 ijimai7_3_00.pdf: 59365 bytes, checksum: 8d7e35f96ee9570c6b66e9776f9b0049 (MD5) https://reunir.unir.net/handle/123456789/13148 https://reunir.unir.net/handle/123456789/13147 Finite Sample Properties of Parameterized Expectations Algorithm Solutions; Is the Length So Determinant? Sánchez-Fuentes, A. Jesús The solution of the Parameterized Expectations Algorithm (PEA) is well defined based on asymptotic properties. In practice, it depends on the specific replication of the exogenous shock(s) used for the resolution process. Typically, this problem is reduced when a sufficiently long replication is considered. In this paper, we suggest an alternative approach which consists of using several, shorter replications. A centrality measure (the median) is used then to discriminate among the different solutions using two different criteria, which differ in the information used. On the one hand, the distance to the vector composed by median values of PEA coefficients is minimized. On the other hand, distances to the median impulse response is minimized. Finally, we explore the impact of considering alternative approaches in an empirical illustration. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-23T07:53:55Z No. of bitstreams: 1 ijimai7_3_3.pdf: 1717721 bytes, checksum: 24b382e746f4eb2114f3cea8942e70e7 (MD5); Made available in DSpace on 2022-05-23T07:53:55Z (GMT). No. of bitstreams: 1 ijimai7_3_3.pdf: 1717721 bytes, checksum: 24b382e746f4eb2114f3cea8942e70e7 (MD5) https://reunir.unir.net/handle/123456789/13147 https://reunir.unir.net/handle/123456789/13140 The Yield Curve as a Recession Leading Indicator. An Application for Gradient Boosting and Random Forest Cadahia Delgado, Pedro; Congregado, Emilio; Golpe, Antonio A.; Vides, José Carlos Most representative decision-tree ensemble methods have been used to examine the variable importance of Treasury term spreads to predict US economic recessions with a balance of generating rules for US economic recession detection. A strategy is proposed for training the classifiers with Treasury term spreads data and the results are compared in order to select the best model for interpretability. We also discuss the use of SHapley Additive exPlanations (SHAP) framework to understand US recession forecasts by analyzing feature importance. Consistently with the existing literature we find the most relevant Treasury term spreads for predicting US economic recession and a methodology for detecting relevant rules for economic recession detection. In this case, the most relevant term spread found is 3-month–6-month, which is proposed to be monitored by economic authorities. Finally, the methodology detected rules with high lift on predicting economic recession that can be used by these entities for this propose. This latter result stands in contrast to a growing body of literature demonstrating that machine learning methods are useful for interpretation comparing many alternative algorithms and we discuss the interpretation for our result and propose further research lines aligned with this work. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T10:35:56Z No. of bitstreams: 1 ijimai7_3_1.pdf: 1003194 bytes, checksum: ac975a0f6cfa3fb863045bc8c7aeed35 (MD5); Made available in DSpace on 2022-05-20T10:35:56Z (GMT). No. of bitstreams: 1 ijimai7_3_1.pdf: 1003194 bytes, checksum: ac975a0f6cfa3fb863045bc8c7aeed35 (MD5) https://reunir.unir.net/handle/123456789/13140 https://reunir.unir.net/handle/123456789/13139 Comparative Analysis of Building Insurance Prediction Using Some Machine Learning Algorithms Ejiyi, Chukwuebuka Joseph; Qin, Zhen; Salako, Abdulhaq Adetunji; Happy, Monday Nkanta; Nneji, Grace Ugochi; Ukwuoma, Chiagoziem Chima; Chikwendu, Ijeoma Amuche; Gen, Ji In finance and management, insurance is a product that tends to reduce or eliminate in totality or partially the loss caused due to different risks. Various factors affect house insurance claims, some of which contribute to formulating insurance policies including specific features that the house has. Machine Learning (ML) when brought into the field of insurance would enable seamless formulation of insurance policies with a better performance which will also save time. Various classification algorithms have been used since they have a long history and have also got some modifications for optimum functionality. To illustrate the performance of each of the ML algorithms that we used here, we analyzed an insurance dataset drawn from Zindi Africa competition which is said to be from Olusola Insurance Company in Lagos Nigeria. This study therefore, compares the performance of Logistic Regression (LR), Decision Tree (DT), K-Nearest Neighbor (KNN), Kernel Support Vector Machine (kSVM), Naïve Bayes (NB), and Random Forest (RF) Regressors on a dataset got from Zindi.africa competition and their performances are checked using not only accuracy and precision metrics but also recall, and F1 score metrics, all displayed on the confusion matrix. The accuracy result shows that logistic regression and Kernel SVM both gave 78% but kSVM outperformed LR in precision with a percentage of 70.8% for kSVM and 64.8% for LR showing that kSVM offered the best result. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T09:53:30Z No. of bitstreams: 1 ijimai7_3_7.pdf: 543039 bytes, checksum: c8857d0579cf9c51fddf8466a5e61ba2 (MD5); Made available in DSpace on 2022-05-20T09:53:30Z (GMT). No. of bitstreams: 1 ijimai7_3_7.pdf: 543039 bytes, checksum: c8857d0579cf9c51fddf8466a5e61ba2 (MD5) https://reunir.unir.net/handle/123456789/13139 https://reunir.unir.net/handle/123456789/13138 An Ensemble Classifier for Stock Trend Prediction Using Sentence-Level Chinese News Sentiment and Technical Indicators Chen, Chun-Hao; Chen, Po-Yeh; Chun-Wei Lin, Jerry In the financial market, predicting stock trends based on stock market news is a challenging task, and researchers are devoted to developing forecasting models. From the existing literature, the performance of the forecasting model is better when news sentiment and technical analysis are considered than when only one of them is used. However, analyzing news sentiment for trend forecasting is a difficult task, especially for Chinese news, because it is unstructured data and extracting the most important features is difficult. Moreover, positive or negative news does not always affect stock prices in a certain way. Therefore, in this paper, we propose an approach to build an ensemble classifier using sentiment in Chinese news at sentence level and technical indicators to predict stock trends. In the training stages, we first divide each news item into a set of sentences. TextRank and word2vec are then used to generate a predefined number of key sentences. The sentiment scores of these key sentences are computed using the given financial lexicon. The sentiment values of the key phrases, the three values of the technical indicators and the stock trend label are merged as a training instance. Based on the sentiment values of the key sets, the corpora are divided into positive and negative news datasets. The two datasets formed are then used to build positive and negative stock trend prediction models using the support vector machine. To increase the reliability of the prediction model, a third classifier is created using the Bollinger Bands. These three classifiers are combined to form an ensemble classifier. In the testing phase, a voting mechanism is used with the trained ensemble classifier to make the final decision based on the trading signals generated by the three classifiers. Finally, experiments were conducted on five years of news and stock prices of one company to show the effectiveness of the proposed approach, and results show that the accuracy and P / L ratio of the proposed approach are 61% and 4.0821 are better than the existing approach. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T09:00:06Z No. of bitstreams: 1 ijimai7_3_5.pdf: 786085 bytes, checksum: d6a9561d04496b63178d717866109231 (MD5); Made available in DSpace on 2022-05-20T09:00:06Z (GMT). No. of bitstreams: 1 ijimai7_3_5.pdf: 786085 bytes, checksum: d6a9561d04496b63178d717866109231 (MD5) https://reunir.unir.net/handle/123456789/13138 https://reunir.unir.net/handle/123456789/13137 AWS PredSpot: Machine Learning for Predicting the Price of Spot Instances in AWS Cloud Baldominos Gómez, Alejandro; Saez, Yago; Quintana, David; Isasi, Pedro Elastic Cloud Compute (EC2) is one of the most well-known services provided by Amazon for provisioning cloud computing resources, also known as instances. Besides the classical on-demand scheme, where users purchase compute capacity at a fixed cost, EC2 supports so-called spot instances, which are offered following a bidding scheme, where users can save up to 90% of the cost of the on-demand instance. EC2 spot instances can be a useful alternative for attaining an important reduction in infrastructure cost, but designing bidding policies can be a difficult task, since bidding under their cost will either prevent users from provisioning instances or losing those that they already own. Towards this extent, accurate forecasting of spot instance prices can be of an outstanding interest for designing working bidding policies. In this paper, we propose the use of different machine learning techniques to estimate the future price of EC2 spot instances. These include linear, ridge and lasso regressions, multilayer perceptrons, K-nearest neighbors, extra trees and random forests. The obtained performance varies significantly between instances types, and root mean squared errors ranges between values very close to zero up to values over 60 in some of the most expensive instances. Still, we can see that for most of the instances, forecasting performance is remarkably good, encouraging further research in this field of study. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T08:30:51Z No. of bitstreams: 1 ijimai7_3_6.pdf: 791834 bytes, checksum: ff2030e80f5bb96f91d3b2fbc0e9feef (MD5); Made available in DSpace on 2022-05-20T08:30:51Z (GMT). No. of bitstreams: 1 ijimai7_3_6.pdf: 791834 bytes, checksum: ff2030e80f5bb96f91d3b2fbc0e9feef (MD5) https://reunir.unir.net/handle/123456789/13137 https://reunir.unir.net/handle/123456789/13136 A Comparative Analysis of Machine Learning Models for Banking News Extraction by Multiclass Classification With Imbalanced Datasets of Financial News: Challenges and Solutions Dogra, Varun; Verma, Sahil; Verma, Kavita; Jhanjhi, NZ; Ghosh, Uttam; Le, Dac-Nhuong Online portals provide an enormous amount of news articles every day. Over the years, numerous studies have concluded that news events have a significant impact on forecasting and interpreting the movement of stock prices. The creation of a framework for storing news-articles and collecting information for specific domains is an important and untested problem for the Indian stock market. When online news portals produce financial news articles about many subjects simultaneously, finding news articles that are important to the specific domain is nontrivial. A critical component of the aforementioned system should, therefore, include one module for extracting and storing news articles, and another module for classifying these text documents into a specific domain(s). In the current study, we have performed extensive experiments to classify the financial news articles into the predefined four classes Banking, Non-Banking, Governmental, and Global. The idea of multi-class classification was to extract the Banking news and its most correlated news articles from the pool of financial news articles scraped from various web news portals. The news articles divided into the mentioned classes were imbalanced. Imbalance data is a big difficulty with most classifier learning algorithms. However, as recent works suggest, class imbalances are not in themselves a problem, and degradation in performance is often correlated with certain variables relevant to data distribution, such as the existence in noisy and ambiguous instances in the adjacent class boundaries. A variety of solutions to addressing data imbalances have been proposed recently, over-sampling, down-sampling, and ensemble approach. We have presented the various challenges that occur with data imbalances in multiclass classification and solutions in dealing with these challenges. The paper has also shown a comparison of the performances of various machine learning models with imbalanced data and data balances using sampling and ensemble techniques. From the result, it’s clear that the performance of Random Forest classifier with data balances using the over-sampling technique SMOTE is best in terms of precision, recall, F-1, and accuracy. From the ensemble classifiers, the Balanced Bagging classifier has shown similar results as of the Random Forest classifier with SMOTE. Random forest classifier's accuracy, however, was 100% and it was 99% with the Balanced Bagging classifier. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T08:12:13Z No. of bitstreams: 1 ijimai7_3_4.pdf: 763331 bytes, checksum: bc3ebf945c8c25485f193bf223598f87 (MD5); Made available in DSpace on 2022-05-20T08:12:13Z (GMT). No. of bitstreams: 1 ijimai7_3_4.pdf: 763331 bytes, checksum: bc3ebf945c8c25485f193bf223598f87 (MD5) https://reunir.unir.net/handle/123456789/13136 https://reunir.unir.net/handle/123456789/13135 Using Customer Knowledge Surveys to Explain Sales of Postgraduate Programs: A Machine Learning Approach Asensio, Eva; Almeida, Alejandro; Galiano, Aida; Martín-Álvarez, Juan Manuel Universities collect information from each customer that contacts them through their websites and social media profiles. Customer knowledge surveys are the main information-gathering tool used to obtain this information about potential students. In this paper, we propose using the information gained via surveys along with enrolment databases, to group customers into homogeneous clusters in order to identify target customers who are more likely to enroll. The use of such a cluster strategy will increase the probability of converting contacts into customers and will allow the marketing and admission departments to focus on those customers with a greater probability of enrolling, thereby increasing efficiency. The specific characteristics of each cluster and those postgraduate programs that are more likely to be selected are identified. In addition, better insight into customers regarding their enrolment choices thanks to our cluster strategy, will allow universities to personalize their services resulting in greater satisfaction and, consequently, in increased future enrolment. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T07:48:35Z No. of bitstreams: 1 ijimai7_3_9.pdf: 745086 bytes, checksum: a0984f3cd37a6e5e718556a73abeaa0d (MD5); Made available in DSpace on 2022-05-20T07:48:35Z (GMT). No. of bitstreams: 1 ijimai7_3_9.pdf: 745086 bytes, checksum: a0984f3cd37a6e5e718556a73abeaa0d (MD5) https://reunir.unir.net/handle/123456789/13135 https://reunir.unir.net/handle/123456789/13134 Re-Evaluating the Relationship Between Economic Development and Self-Employment, at the Macro-Level: A Bayesian Model Averaging Approach Rodriguez-Santiago, Ana We re-evaluate the relationship between stages of economic development and entrepreneurship, at the macro level. We first conduct a literature review of previous empirical research on cross-country determinants of entrepreneurship in order to put our contribution in perspective. To circumvent problems related to model uncertainty we use Bayesian Model Averaging (BMA) to evaluate the robustness of determinants of economic growth in a new dataset of 117 countries in the 2005-2019 period, allowing fixed effects and investigating the existence of heterogeneity allowing interactions of our focus variable with other regressors. Our empirical analysis then shows that the variation of self-employment rates across countries are mainly determined by variations in the unemployment, the stage of economic development and the variations in labor market frictions. When interactions are taken into account, results confirm that there is a differential effect of labor market frictions in countries with different levels of income. Frictions in labor market may encourage becoming self-employed in richer countries. Submitted by Susana Figueroa Navarro (susana.figueroa.n@unir.net) on 2022-05-20T07:20:53Z No. of bitstreams: 1 ijimai7_3_2.pdf: 672659 bytes, checksum: 8bc024d1f11be3e9c2717950ddc43015 (MD5); Made available in DSpace on 2022-05-20T07:20:53Z (GMT). No. of bitstreams: 1 ijimai7_3_2.pdf: 672659 bytes, checksum: 8bc024d1f11be3e9c2717950ddc43015 (MD5) https://reunir.unir.net/handle/123456789/13134