[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119496-en":3,"doc-seo-119496-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119496,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","Preventing university drop outs - Early detection of university drop outs using machine learning - Master’s Thesis","University dropout rates create major pressure on higher education institutions by reducing graduation outcomes and wasting resources. This Master’s thesis investigates machine learning as an early detection approach to identify students at risk of dropping out. Using academic data from SISU, a Finnish student information system, multiple models were tested and optimized. Key steps include feature engineering, SMOTE-based data balancing, and evaluation to support reliable predictions. Results show academic data alone can build a viable predictive system, enabling data-driven interventions to improve student retention.","Preventing university drop outs  \nEarly detection of university drop outs using machine learning  \nEddie Fernberg  \nMaster’s Thesis  \nMEng in Big Data Analytics 2025  \nMaster’s Thesis  \nEddie Fernberg  \nPreventing university drop outs. Early detection of university drop outs using machine learning.  \nArcada University of Applied Sciences: MEng in Big Data Analytics, 2025 .  \nIdentification number:  \n048147433  \nCommissioned by:  \nN/A  \nAbstract:  \nUniversity drop out rates present a significant challenge for higher education institutions, leading to wasted resources and decreased graduation rates. This thesis explores the potential of machine learning to provide an early detection system for identifying students at risk of dropping out. Using academic data from SISU, a student information system used in Finland, various machine learning models were tested and optimized. Key steps included feature engineering, data balancing using SMOTE, and model evaluation to ensure reliable predictions. The results indicate that academic data alone is sufficient for creating a viable predictive system. The findings emphasize the importance of early intervention strategies and the potential for data-driven decision-making in university administration. This study demonstrates that machine learning can serve as a powerful tool for drop out prevention, enabling universities to take proactive steps toward student retention.  \nKeywords: Dropout detection, machine learning, SISU, feature engineering  \nContents  \n1 Introduction .................................. 8  \n1.1 Background .................................. 9  \n2 Related Work .................................. 10  \n2.1 Literature study ................................ 10  \n2.1.1 Academic vs. personal data ........................ 11  \n2.1.2 Data sources ............................... 13  \n2.1.3 Balancing ................................ 13  \n2.1.4 Models .................................. 14  \n3 Research Methodology ............................ 15  \n3.1 Tools ..................................... 15  \n3.2 Acquiring data ................................ 17  \n3.3 Dataset .................................... 17  \n3.4 Data preparation ............................... 18  \n3.4.1 Filtering ................................. 18  \n3.5 Feature engineering .............................. 19  \n3.5.1 Dependent variables ........................... 19  \n3.5.2 Column completeness .......................... 21  \n3.5.3 Feature amount .............................. 21  \n3.5.4 Imbalances ................................ 23  \n4 Experiments .................................. 25  \n4.1 Models .................................... 25  \n4.2 Predictions .................................. 26  \n4.3 Predictions with SMOTE ............................ 32  \n5 Results ..................................... 38  \n6 Conclusions .................................. 39  \n6.0.1 Findings ................................. 39  \n6.0.2 Further Research ............................. 39  \nReferences ...................................... 41  \nAppendix A ...................................... 45  \nAppendix B ...................................... 48  \nAppendix C ...................................... 51  \nFigures  \nFigure 1 . Correlation matrix at data load ...................... 19  \nFigure 2 . Column completeness after drop and dropna ............... 22  \nFigure 3 . Correlation matrix after drop and dropna ................. 23  \nFigure 4 . Drop out status distribution ........................ 24  \nFigure 5 . Drop out status distribution for test set .................. 27  \nFigure 6 . ROC curves for models .......................... 29  \nFigure 7 . Confusion matrices ............................ 30  \nFigure 8 . Feature importances ........................... 31  \nFigure 9 . Drop out status distribution for test set (after SMOTE) ......... 33  \nFigure 10 . ROC curves for models after SMOTE .................. 34  \nFigure 11 . Confusion matric","cbCail0kMfbp4byY","https://ap.wps.com/l/cbCail0kMfbp4byY","pdf",2073891,1,52,"English","en",105,"# Introduction\n## Background\n# Related Work\n## Literature study\n### Academic vs. personal data\n### Data sources\n### Balancing\n### Models\n# Research Methodology\n## Tools\n## Acquiring data\n## Dataset\n## Data preparation\n### Filtering\n## Feature engineering\n### Dependent variables\n### Column completeness\n### Feature amount\n### Imbalances\n# Experiments\n## Models\n## Predictions\n## Predictions with SMOTE\n# Results\n# Conclusions\n## Findings\n## Further Research\n# References\n# Appendix A\n# Appendix B\n# Appendix C","[{\"question\":\"What problem does the thesis address?\",\"answer\":\"The thesis addresses the challenge of university dropout rates and the need to reduce wasted resources and lower graduation outcomes through earlier identification of at-risk students.\"},{\"question\":\"Which data source and methods are used for early detection?\",\"answer\":\"Academic data from SISU, Finland’s student information system, is used. The approach includes feature engineering, SMOTE for data balancing, and model evaluation to produce reliable dropout-risk predictions.\"},{\"question\":\"Do academic data alone provide sufficient predictive power?\",\"answer\":\"Yes. The results indicate that academic data alone is sufficient to create a viable predictive system for dropout detection.\"}]","Preventing university drop outs - Early detection of university drop outs using machine learning - Master’s Thesis | PDF",1785724621,131,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"preventing-university-drop-outs-early-detection-of-university-drop-outs-using-machine-learning-masters-thesis","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/preventing-university-drop-outs-early-detection-of-university-drop-outs-using-machine-learning-masters-thesis/119496/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the thesis address?","Question",{"text":75,"@type":76},"The thesis addresses the challenge of university dropout rates and the need to reduce wasted resources and lower graduation outcomes through earlier identification of at-risk students.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which data source and methods are used for early detection?",{"text":80,"@type":76},"Academic data from SISU, Finland’s student information system, is used. The approach includes feature engineering, SMOTE for data balancing, and model evaluation to produce reliable dropout-risk predictions.",{"name":82,"@type":73,"acceptedAnswer":83},"Do academic data alone provide sufficient predictive power?",{"text":84,"@type":76},"Yes. The results indicate that academic data alone is sufficient to create a viable predictive system for dropout detection.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]