Data Science and Management最新文献

{"title":"Effective and efficient handling of missing data in supervised machine learning","authors":"Peter Ayokunle Popoola ,&nbsp;Jules-Raymond Tapamo ,&nbsp;Alain Guy Honoré Assounga","doi":"10.1016/j.dsm.2024.12.002","DOIUrl":"10.1016/j.dsm.2024.12.002","url":null,"abstract":"<div><div>The prevailing consensus in statistical literature is that multiple imputation is generally the most suitable method for addressing missing data in statistical analyses, whereas a complete case analysis is deemed appropriate only when the rate of missingness is negligible or when the missingness mechanism is missing completely at random (MCAR). This study investigates the applicability of this consensus within the context of supervised machine learning, with particular emphasis on the interactions between the imputation method, missingness mechanism, and missingness rate. Furthermore, we examine the time efficiency of these “state-of-the-art” imputation methods considering the time-sensitive nature of certain machine learning applications. Utilizing ten real-world datasets, we introduced missingness at rates ranging from approximately 5%–75% under the MCAR, missing at random (MAR), and missing not at random (MNAR) mechanisms. We subsequently address missing data using five methods: complete case analysis (CCA), mean imputation, hot deck imputation, regression imputation, and multiple imputation (MI). Statistical tests are conducted on the machine learning outcomes, and the findings are presented and analyzed. Our investigation reveals that in nearly all scenarios, CCA performs comparably to MI, even with substantial levels of missingness under the MAR and MNAR conditions and with missingness in the output variable for regression problems. Under some conditions, CCA surpasses MI in terms of its performance. Thus, given the considerable computational demands associated with MI, the application of CCA is recommended within the broader context of supervised machine learning, particularly in big-data environments.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 3","pages":"Pages 361-373"},"PeriodicalIF":0.0,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144931740","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Meta-model-based optimization of rule-based energy management in second-hand plug-in hybrid electric vehicles","authors":"Debraj Bhattacharjee ,&nbsp;Sourabh Mandol ,&nbsp;Tamal Ghosh","doi":"10.1016/j.dsm.2024.12.003","DOIUrl":"10.1016/j.dsm.2024.12.003","url":null,"abstract":"<div><div>This study presents a methodology to enhance energy management systems (EMS) in hybrid electric vehicles (HEVs) to reduce fuel consumption and greenhouse gas emissions. A novel surrogate-assisted optimization framework is employed, incorporating key performance metrics such as fuel efficiency and emissions to develop data-driven surrogate models of the EMS. These models are optimized using various algorithms targeting parameters such as engine idle speed, thermostat temperature fraction, regeneration load factor, and battery state-of-charge thresholds. Correlation analysis highlights the significant impact of the lower state-of-charge threshold and thermostat temperature fraction on fuel efficiency and emissions. Among the optimization methods, the combination of a backpropagation neural network (BPNN) and a multi-objective genetic algorithm (MOGA) proves most effective, achieving fuel consumption reductions of 5.26% and 5.01% in charge-sustaining and charge-depletion modes, respectively. Additionally, the BPNN-based MOGA demonstrates notable improvements in emission reduction. These findings suggest that optimizing rule-based EMS parameters without altering underlying management rules can significantly enhance performance under diverse and unanticipated driving conditions.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 3","pages":"Pages 388-402"},"PeriodicalIF":0.0,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145019248","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Electroencephalogram-based emotion recognition: a comparative analysis of supervised machine learning algorithms","authors":"Anagha Prakash ,&nbsp;Alwin Poulose","doi":"10.1016/j.dsm.2024.12.004","DOIUrl":"10.1016/j.dsm.2024.12.004","url":null,"abstract":"<div><div>Emotion recognition from electroencephalogram (EEG) signals has garnered significant attention owing to its potential applications in affective computing, human-computer interaction, and mental health monitoring. This paper presents a comparative analysis of different machine learning methods for emotion recognition using EEG data. The objective of this study was to identify the most effective algorithm for accurately classifying emotional states using EEG signals. The EEG brainwave dataset: Feeling emotions dataset was used to evaluate the performance of various machine-learning techniques. Multiple machine learning techniques, namely logistic regression (LR), support vector machine (SVM), Gaussian Naive Bayes (GNB), and decision tree (DT), and ensemble models, namely random forest (RF), AdaBoost, LightGBM, XGBoost, and CatBoost, were trained and evaluated. Five-fold cross-validation and dimension reduction techniques, such as principal component analysis, <em>t</em>-distributed stochastic neighbor embedding, and linear discriminant analysis, were performed for all models. The least-performing model, GNB, showed substantially increased performance after dimension reduction. Performance metrics such as accuracy, precision, recall, F1-score, and receiver operating characteristic curves are employed to assess the effectiveness of each approach. This study focuses on the implications of using various machine learning algorithms for EEG-based emotion recognition. This pursuit can improve our understanding of emotions and their underlying neural mechanisms.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 3","pages":"Pages 342-360"},"PeriodicalIF":0.0,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144931739","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Toward accurate credit evaluation: an efficient imputation approach for financial data","authors":"Jie Lu ,&nbsp;Shengda Zhuo ,&nbsp;Jinjie Qiu ,&nbsp;Yin Tang","doi":"10.1016/j.dsm.2025.06.001","DOIUrl":"10.1016/j.dsm.2025.06.001","url":null,"abstract":"<div><div>Missing instances and mixed data types, including discrete and ordered (e.g., continuous and ordinal) variables, are widespread in many datasets in the finance sector. In this domain, estimating missing instances is crucial because many data analysis pipelines require complete data, which is particularly challenging for mixed-type data. However, existing methods treat discrete and ordinal data as continuous values, which may reduce efficacy in addressing these challenges. To fill this gap, this study proposes a probabilistic imputation method for mixed-type and incomplete loan data (PMILD), using a mixed Gaussian Copula model that supports single and multiple imputations. The method models mixed discrete and ordinal data using latent Gaussian distributions, where observed features with arbitrary margins are mapped to the latent normal space, and feature correlations are approximated through the expectation-maximization process in the latent space. Empirical results on nine real-world datasets demonstrate that PMILD substantially outperforms state-of-the-art imputation methods, providing a highly effective solution for handling mixed-type and incomplete loan data. This advancement enhances both operational efficiency and credit evaluation accuracy in finance-related applications.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 3","pages":"Pages 374-387"},"PeriodicalIF":0.0,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144988502","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Categorical classification of skin cancer using a weighted ensemble of transfer learning with test time augmentation","authors":"Aliyu Tetengi Ibrahim ,&nbsp;Mohammed Abdullahi ,&nbsp;Armand Florentin Donfack Kana ,&nbsp;Mohammed Tukur Mohammed ,&nbsp;Ibrahim Hayatu Hassan","doi":"10.1016/j.dsm.2024.10.002","DOIUrl":"10.1016/j.dsm.2024.10.002","url":null,"abstract":"<div><div>Skin cancer is the abnormal development of cells on the surface of the skin and is one of the most fatal diseases in humans. It usually appears in locations that are exposed to the sun, but can also appear in areas that are not regularly exposed to the sun. Due to the striking similarities between benign and malignant lesions, skin cancer detection remains a problem, even for expert dermatologists. Considering the inability of dermatologists to diagnose skin cancer accurately, a convolutional neural network (CNN) approach was used for skin cancer diagnosis. However, the CNN model requires a significant number of image datasets for better performance; thus, image augmentation and transfer learning techniques have been used in this study to boost the number of images and the performance of the model, because there are a limited number of medical images. This study proposes an ensemble transfer-learning-based model that can efficiently classify skin lesions into one of seven categories to aid dermatologists in skin cancer detection: (i) actinic keratoses, (ii) basal cell carcinoma, (iii) benign keratosis, (iv) dermatofibroma, (v) melanocytic nevi, (vi) melanoma, and (vii) vascular skin lesions. Five transfer learning models were used as the basis of the ensemble: MobileNet, EfficientNetV2B2, Xception, ResNext101, and DenseNet201. In addition to the stratified 10-fold cross-validation, the results of each individual model were fused to achieve greater classification accuracy. An annealing learning rate scheduler and test time augmentation (TTA) were also used to increase the performance of the model during the training and testing stages. A total of 10,015 publicly available dermoscopy images from the HAM10000 (Human Against Machine) dataset, which contained samples from the seven common skin lesion categories, were used to train and evaluate the models. The proposed technique attained 94.49% accuracy on the dataset. These results suggest that this strategy can be useful for improving the accuracy of skin cancer classification. However, the weighted average of F1-score, recall, and precision were obtained to be 94.68%, 94.49%, and 95.07%, respectively.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 174-184"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144222281","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Challenges and prospects of artificial intelligence in aviation: a ​bibliometric study","authors":"Nuno Moura Lopes,&nbsp;Manuela Aparicio,&nbsp;Fátima Trindade Neves","doi":"10.1016/j.dsm.2024.11.001","DOIUrl":"10.1016/j.dsm.2024.11.001","url":null,"abstract":"<div><div>The primary motivation for this study is the recent growth and increased interest in artificial intelligence (AI). Despite the widespread recognition of its critical importance, a discernible scientific gap persists within the extant scholarly discourse, particularly concerning exhaustive systematic reviews of AI in the aviation industry. This gap spurred a meticulous analysis of 1,213 articles from the Web of Science (WoS) core database for bibliometric knowledge mapping. This analysis highlights China as the primary contributor to publications, with the Nanjing University of Finance and Economics as the leading institution in paper contributions. <em>Lecture Notes in Artificial Intelligence</em> and the <em>IEEE AIAA Digital Avionics System Conference</em> are the leading journals within this domain. This bibliometric research underscores the key focus on air traffic management, human factors, environmental initiatives, training, logistics, flight operations, and safety through co-occurrence and co-citation analyses. A chronological examination of keywords reveals a central research trajectory centered on machine learning, models, deep learning, and the impact of automation on human performance in aviation. Burst keyword analysis identifies the leading-edge research on AI within predictive models, unmanned aerial vehicles, object detection, and convolutional neural networks. The primary objective is to bridge this knowledge gap and gain comprehensive insights into AI in the aviation sector. This study delineates the scholarly terrain of AI in aviation using a bibliometric methodology to facilitate this exploration. The results illuminate the current state of research, thereby enhancing academic understanding of developments within this critical domain. Finally, a new conceptual framework was constructed based on the primary elements identified in the literature. This framework can assist emerging researchers in identifying the fundamental dimensions of AI in the aviation industry.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 207-223"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144263658","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel method for a technology enhanced learning recommender system considering changing user interest based on neural collaborative filtering","authors":"Mohammad Mehran Lesan Sedgh ,&nbsp;Alimohammad Latif ,&nbsp;Sima Emadi","doi":"10.1016/j.dsm.2024.09.004","DOIUrl":"10.1016/j.dsm.2024.09.004","url":null,"abstract":"<div><div>This study introduces an advanced recommender system for technology enhanced learning (TEL) that synergizes neural collaborative filtering, sentiment analysis, and an adaptive learning rate to address the limitations of traditional TEL systems. Recognizing the critical gap in existing approaches—primarily their neglect of user emotional feedback and static learning paths—our model innovatively incorporates sentiment analysis to capture and respond to nuanced emotional feedback from users. Utilizing bidirectional encoder representations from Transformers for sentiment analysis, our system not only understands but also respects user privacy by processing feedback without revealing sensitive information. The adaptive learning rate, inspired by AdaGrad, allows our model to adjust its learning trajectory based on the sentiment scores associated with user feedback, ensuring a dynamic response to both positive and negative sentiments. This dual approach enhances the system’s adaptability to changing user preferences and improves its contentment understanding. Our methodology involves a comprehensive analysis of both the content of learning materials and the behaviors and preferences of learners, facilitating a more personalized learning experience. By dynamically adjusting recommendations based on real-time user data and behavioral analysis, our system leverages the collective insights of similar users and relevant content. We validated our approach against three datasets—MovieLens, Amazon, and a proprietary TEL dataset—and saw significant improvements in recommendation precision, F-score, and mean absolute error. The results indicate the potential of integrating sentiment analysis and adaptive learning rates into TEL recommender systems, marking a step forward in developing more responsive and user-centric educational technologies. This study paves the way for future advancements in TEL systems, emphasizing the importance of emotional intelligence and adaptability in enhancing the learning experience.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 196-206"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144255255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Factors influencing readiness for artificial intelligence: a systematic literature review","authors":"Wajid Ali,&nbsp;Abdul Zahid Khan","doi":"10.1016/j.dsm.2024.09.005","DOIUrl":"10.1016/j.dsm.2024.09.005","url":null,"abstract":"<div><div>Public-and private-sector organizations have adopted artificial intelligence (AI) to meet the challenges of the Fourth Industrial Revolution. The successful implementation of AI is a challenging task, and previous research has advocated the need to explore key readiness before AI implementation. The objective of this study is to identify the AI readiness factors explored by different authors in past research. To achieve this, we conducted a rigorous literature review. The approach used in the systematic literature review is also discussed. A rigorous review of 52 studies from various journals and databases (Science Direct, Springer Link, Institute of Electrical and Electronics Engineers, Emerald, and Google Scholar) identified 23 AI readiness factors. The key factors identified were mainly related to organizational information technology infrastructure, top management support, resource availability, collaborative culture, organizational size, organizational capability, compatibility, data quality, and financial budget, whereas the other 15 were potential factors in AI readiness. All of these factors should be considered before the implementation of AI in any organization. The findings also reflect a high failure rate, including AI readiness factors, which are intended to facilitate AI adoption in organizations and reduce the frequency of failures. These factors will aid management in developing an effective strategy for AI implementation in organizations.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 224-236"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144263659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Surprisal-based algorithm for detecting anomalies in categorical data","authors":"Ossama Cherkaoui,&nbsp;Houda Anoun,&nbsp;Abderrahim Maizate","doi":"10.1016/j.dsm.2025.01.005","DOIUrl":"10.1016/j.dsm.2025.01.005","url":null,"abstract":"<div><div>Anomaly detection is an important research area in a diverse range of real-world applications. Although many algorithms have been proposed to address anomaly detection for numerical datasets, categorical and mixed datasets remain a significant challenge, primarily because a natural distance metric is lacking. Consequently, the methods proposed in the literature implement entirely different assumptions regarding the definition of categorical anomalies. This paper presents a novel categorical anomaly detection approach, offering two key contributions to existing methods. First, a novel surprisal-based anomaly score is introduced, which provides a more accurate assessment of anomalies by considering the full distribution of categorical values. Second, the proposed method considers complex correlations in the data beyond the pairwise interactions of features. This study proposed and tested the novel categorical surprisal anomaly detection algorithm (CSAD) by comparing and evaluating it against six competitors. The experimental results indicate that CSAD produced the best overall performance, achieving the highest average ROC-AUC and PR-AUC values of 0.8 and 0.443, respectively. Furthermore, CSAD's execution time is satisfactory even when processing large, high-dimensional datasets.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 185-195"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144222282","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Triadic concept analysis for insights extraction from longitudinal studies in health","authors":"João Pedro Santos,&nbsp;Atílio Ferreira Silva,&nbsp;Henrique Fernandes Viana Mendes,&nbsp;Mark Alan Junho Song,&nbsp;Luis Enrique Zárate","doi":"10.1016/j.dsm.2024.10.001","DOIUrl":"10.1016/j.dsm.2024.10.001","url":null,"abstract":"<div><div>In the health field, longitudinal studies involve the recording of clinical observations of the same sample of patients over successive periods, referred to as waves. This type of database serves as a valuable source of information and insights, particularly when examining the temporal aspect, allowing the extraction of relevant and non-obvious knowledge. The triadic concept analysis theory has been proposed to describe the ternary relationships between objects, attributes, and conditions. In this study, we present a methodology for exploring longitudinal health databases using both the triadic theory and triadic rules, which are similar to association rules but incorporate temporal relations. Through four case studies, we demonstrate the potential of applying triadic analysis to longitudinal databases to identify risk patterns, enhance decision-making processes, and deepen our understanding of temporal dynamics. These findings suggest a promising approach for describing longitudinal databases and obtaining insights to improve clinical decision-support systems for disease treatment.</div></div>","PeriodicalId":100353,"journal":{"name":"Data Science and Management","volume":"8 2","pages":"Pages 160-173"},"PeriodicalIF":0.0,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144222283","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}