[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-id-113":3,"doc-seo-128393-113":53,"doc-detail-128393-id":128},{"code":4,"msg":5,"data":6},0,"success",[7,13,17,21,25,29,33,37,41,45,49],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},55,"Document","Agama & Spiritualitas",60,"religion-spirituality",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":11,"slug":16},48,"Cerita & Novel","story-novel",{"id":18,"doc_module":4,"doc_module_name":9,"category_name":19,"show_sort_weight":11,"slug":20},56,"Gaya Hidup","lifestyle",{"id":22,"doc_module":4,"doc_module_name":9,"category_name":23,"show_sort_weight":11,"slug":24},51,"Komik","comic",{"id":26,"doc_module":4,"doc_module_name":9,"category_name":27,"show_sort_weight":11,"slug":28},53,"Layanan Kesehatan","healthcare",{"id":30,"doc_module":4,"doc_module_name":9,"category_name":31,"show_sort_weight":11,"slug":32},54,"Penelitian & Laporan","research-report",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":11,"slug":36},49,"Sastra","literature",{"id":38,"doc_module":4,"doc_module_name":9,"category_name":39,"show_sort_weight":11,"slug":40},52,"Teknologi","technology",{"id":42,"doc_module":4,"doc_module_name":9,"category_name":43,"show_sort_weight":11,"slug":44},50,"Ujian","exam",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":47,"show_sort_weight":11,"slug":48},57,"Umum","general",{"id":50,"doc_module":4,"doc_module_name":9,"category_name":51,"show_sort_weight":4,"slug":52},181,"Formulir","formulir",{"code":4,"msg":54,"data":55},"ok",{"site_id":56,"language":57,"slug":58,"title":59,"keywords":60,"description":61,"schema_data":62,"social_meta":121,"head_meta":123,"extra_data":125,"updated_unix":127},113,"id","development-of-an-early-detection-system-for-at-risk-students-using-data-based-machine-learning-case-study","Pengembangan Sistem Deteksi Dini Mahasiswa Berisiko Menggunakan Machine Learning Berbasis Data Learning Management System - Studi Kasus","","Penelitian ini mengembangkan sistem deteksi dini mahasiswa berisiko menggunakan machine learning berbasis data dari Learning Management System (LMS) rumahilmu.org. Sistem dirancang untuk Program Studi Sistem Informasi Universitas Bengkulu dengan menganalisis data 459 enrollment pada lima mata kuliah. Sebanyak 37–76 fitur diekstraksi dari aktivitas LMS untuk memprediksi mahasiswa berpotensi memperoleh nilai di bawah persentil ke-30 pada 25%, 50%, dan 75% semester. Optimasi per-kelas diuji pada 11 algoritma; tidak ada algoritma tunggal unggul, sedangkan recall kelas minoritas 'Berisiko' sangat rendah (Recall=0.00 pada 8/15 skenario), meski F1-score validasi silang >0.80 mengindikasikan overfitting. Arsitektur 3-tier (FastAPI dan React) menyediakan dashboard interaktif namun efektivitas dibatasi dataset kecil dan tidak seimbang.",{"@graph":63,"@context":120},[64,81,103],{"@type":65,"itemListElement":66},"BreadcrumbList",[67,72,75,78],{"item":68,"name":69,"@type":70,"position":71},"https://docshare.wps.com","Home","ListItem",1,{"item":73,"name":9,"@type":70,"position":74},"https://docshare.wps.com/id/document/",2,{"item":76,"name":31,"@type":70,"position":77},"https://docshare.wps.com/id/document/penelitian-laporan/",3,{"item":79,"name":59,"@type":70,"position":80},"https://docshare.wps.com/id/document/development-of-an-early-detection-system-for-at-risk-students-using-data-based-machine-learning-case-study/128393/",4,{"url":79,"name":59,"@type":82,"image":83,"author":88,"headline":59,"publisher":91,"fileFormat":94,"inLanguage":57,"description":61,"dateModified":95,"datePublished":96,"encodingFormat":94,"isAccessibleForFree":97,"interactionStatistic":98},"DigitalDocument",{"url":84,"@type":85,"width":86,"height":87},"https://docshare.wps.com/thumbnails/development-of-an-early-detection-system-for-at-risk-students-using-data-based-machine-learning-case-study/128393.png","ImageObject",300,407,{"name":89,"@type":90},"Aurelia","Person",{"url":68,"name":92,"@type":93},"DocShare","Organization","application/pdf","2026-09-19","2026-08-05",true,{"@type":99,"interactionType":100,"userInteractionCount":102},"InteractionCounter",{"@type":101},"ViewAction",7,{"@type":104,"mainEntity":105},"FAQPage",[106,112,116],{"name":107,"@type":108,"acceptedAnswer":109},"Bagaimana sistem mendeteksi mahasiswa berisiko dalam penelitian ini?","Question",{"text":110,"@type":111},"Sistem memanfaatkan data aktivitas dari LMS rumahilmu.org, mengekstraksi 37–76 fitur, lalu memakai machine learning untuk memprediksi mahasiswa yang berpotensi mendapat nilai di bawah persentil ke-30 pada titik waktu 25%, 50%, dan 75% semester.","Answer",{"name":113,"@type":108,"acceptedAnswer":114},"Apakah ada satu algoritma yang paling unggul untuk semua mata kuliah?",{"text":115,"@type":111},"Tidak. Hasil per-kelas menunjukkan tidak ada algoritma tunggal yang superior untuk semua mata kuliah. Model paling efektif berbeda-beda, dengan Gaussian Process, Logistic Regression, dan Voting Classifier sering terpilih.",{"name":117,"@type":108,"acceptedAnswer":118},"Apa kendala utama yang ditemukan saat evaluasi model?",{"text":119,"@type":111},"Evaluasi mengindikasikan overfitting dan penurunan performa walaupun validasi silang memiliki F1-score tinggi (lebih dari 0.80). Kemampuan mendeteksi kelas minoritas 'Berisiko' juga rendah, dengan Recall 0.00 pada 8 dari 15 skenario.","https://schema.org",{"og:url":79,"og:type":122,"og:title":59,"og:site_name":92,"og:description":61},"article",{"robots":124,"canonical":79},"index,follow",{"doc_id":126,"site_id":56},128393,1785947269,{"code":4,"msg":5,"data":129},{"doc_id":126,"user_id":130,"nickname":89,"user_avatar":131,"doc_module":4,"category_id":30,"category_name":31,"doc_title":59,"doc_description":61,"doc_content":132,"file_id":133,"file_url":134,"file_type":135,"file_size":136,"view_count":102,"is_deleted":4,"is_public":71,"is_downloadable":71,"audit_status":71,"page_count":137,"language":138,"language_code":57,"site_id":56,"html_lang":57,"table_of_contents":139,"faqs":140,"seo_title":141,"seo_description":61,"update_tm":127,"read_time":142},962085564807,"https://ap-avatar.wpscdn.com/davatar_6f874abed73319feea01a86fa6f0fab8","PENGEMBANGAN SISTEM DETEKSI DINI MAHASISWA BERISIKO MENGGUNAKAN MACHINE LEARNING BERBASIS DATA LEARNING MANAGEMENTSYSTEM  \n(Studi Kasus: [rumahilmu.org](rumahilmu.org))  \nWahyu Syahputra 1, Endina Putri Purwandari2, Widhia KZ Oktoeberza3  \n1,2,,3 Program Studi Informatika, Fakultas Teknik, Universitas Bengkulu,  \nJl. WR. Supratman, Kandang Limun, Bengkulu, 3871A  \nTelp. (627) 3621170, Faks (627) 3622105  \n[1](1syahputrawahyu61@gmail.com)[syahputrawahyu61@gmail.com](1syahputrawahyu61@gmail.com)  \n[2](2endinaputri@unib.ac.id)[endinaputri@unib.ac.id](2endinaputri@unib.ac.id)  \n[3](3widhiakz@unib.ac.id)[widhiakz@unib.ac.id](3widhiakz@unib.ac.id)  \nAbstrak: Penelitian ini bertujuan mengembangkan sistem deteksi dini mahasiswa berisiko menggunakan machine learning berbasis data dari Learning Management System (LMS) [rumahilmu.org](rumahilmu.org). Sistem ini dirancang untuk Program Studi Sistem Informasi Universitas Bengkulu, dengan menganalisis data dari 459 enrollment mahasiswa pada lima mata kuliah. Sebanyak 37–76 fitur diekstraksi dariaktivitas LMS untuk memprediksi mahasiswa yang berpotensi mendapat nilai di bawah persentil ke-30 pada tiga titik waktu strategis (25%, 50%, dan 75% semester) . Penelitian ini menerapkan pendekatan optimasi per-kelas, menguji 11 algoritma untuk menemukan model terbaik bagi setiap mata kuliah. Hasil penelitian menunjukkan bahwa tidak ada satu algoritma tunggal yang superior; model paling efektif bervariasi untuk setiap mata kuliah, dengan Gaussian Process, Logistic Regression, dan Voting Classifier menjadi yang paling sering terpilih. Namun, evaluasi pada data uji menunjukkan tantangan signifikan: meskipun skor validasi silang tinggi (F1-score > 0.80), terjadi overfitting dan penurunan performa. Temuan paling krusial adalah rendahnya kemampuan model dalam mendeteksi kelas minoritas 'Berisiko', dengan metrik Recall (Berisiko) mencapai 0.00 pada 8 dari 15 skenario. Kinerja deteksi terbaik dicapaipada mata kuliah Statistika & Probabilitas dengan Recall 0.50. Sistem yang diimplementasikan denganarsitektur 3-tier (FastAPI dan React) menyediakan dashboard interaktif, namun efektivitas prediksinyauntuk deteksi dini dibatasi oleh dataset yang kecil dan tidak seimbang.  \nKata Kunci: deteksi dini mahasiswa berisiko, learning analytics, machine learning, klasifikasi biner, learning management system  \nAbstract: This research aims to develop an early detection system for at-risk students using machine learning based on data from the Learning Management System (LMS) [rumahilmu.org. The system was](rumahilmu.org. The system was) designed for the Information Systems Study Programs at the University of Bengkulu, analyzing data from 459 student enrollments across five courses. A total of 37–76 features were extracted from LMS activities to predict students likely to score below the 30th percentile at three strategic time points (25%, 50%, and 75% of the semester). This study implemented a per-class optimization approach, testing 11 algorithms to find the best model for each course. The results showed that no single algorithm was universally superior; the most effective models varied for each course, with Gaussian Process, Logistic Regression, and Voting Classifier being the most frequently chosen. However, evaluation on the test data revealed significant challenges: despite highcross-validation scores (F1-score > 0.80), overfitting and performance degradation occurred. The most critical finding was the model's low capability in detecting the 'At-Risk'minority class, with the Recall (At-Risk) metric reaching 0.00 in 8 out of 15 scenarios. The best detection performance was achieved in the Statistics & Probability course with a Recall of 0.50. The implemented system, featuring a 3-tier architecture (FastAPI and React), provides an interactive dashboard, but its predictive effectiveness for early detection is limited by small and imbalanced datasets.  \nKeywords: early detection of at-risk students, learning analytics, ma","cbCaiibtEVeycYPh","https://ap.wps.com/l/cbCaiibtEVeycYPh","pdf",887216,15,"Indonesian","# Pendahuluan\n## Latar belakang pembelajaran konvensional di perguruan tinggi\n## Dampak era digital terhadap dinamika kelas\n## Kebutuhan transformasi pembelajaran personal dan berbasis data","[{\"question\":\"Bagaimana sistem mendeteksi mahasiswa berisiko dalam penelitian ini?\",\"answer\":\"Sistem memanfaatkan data aktivitas dari LMS rumahilmu.org, mengekstraksi 37–76 fitur, lalu memakai machine learning untuk memprediksi mahasiswa yang berpotensi mendapat nilai di bawah persentil ke-30 pada titik waktu 25%, 50%, dan 75% semester.\"},{\"question\":\"Apakah ada satu algoritma yang paling unggul untuk semua mata kuliah?\",\"answer\":\"Tidak. Hasil per-kelas menunjukkan tidak ada algoritma tunggal yang superior untuk semua mata kuliah. Model paling efektif berbeda-beda, dengan Gaussian Process, Logistic Regression, dan Voting Classifier sering terpilih.\"},{\"question\":\"Apa kendala utama yang ditemukan saat evaluasi model?\",\"answer\":\"Evaluasi mengindikasikan overfitting dan penurunan performa walaupun validasi silang memiliki F1-score tinggi (lebih dari 0.80). Kemampuan mendeteksi kelas minoritas 'Berisiko' juga rendah, dengan Recall 0.00 pada 8 dari 15 skenario.\"}]","Pengembangan Sistem Deteksi Dini Mahasiswa Berisiko Menggunakan Machine Learning Berbasis Data Learning Management System - Studi Kasus | PDF",23]