NURYANA, I Kadek Dwi and Purwanto, Purwanto and Isnanto, R. Rizal (2025) MODEL INTERAKSI MAHASISWA PADA KINERJA AKADEMIK DALAM PEMBELAJARAN DARING MENGGUNAKAN STRUCTURAL EQUATION MODELING PARTIAL LEAST SQUARES (SEM-PLS) DAN RANDOM FOREST REGRESSOR. Doctoral thesis, UNIVERSITAS DIPONEGORO.

	Text cover aja.pdf Download (211kB)
	Text COVER LENGKAP..pdf Restricted to Repository staff only Download (2MB)
	Text BAB I..pdf Download (513kB)
	Text BAB II..pdf Download (1MB)
	Text BAB III..pdf Restricted to Repository staff only Download (684kB)
	Text BAB IV..pdf Restricted to Repository staff only Download (1MB)
	Text BAB V..pdf Restricted to Repository staff only Download (387kB)
	Text DAFTAR REFERENSI..pdf Download (306kB)
	Text LAMPIRAN DISERTASI..pdf Restricted to Repository staff only Download (3MB)

Abstract

Perkembangan pembelajaran daring telah membawa perubahan signifikan dalam sistem pendidikan tinggi, namun juga menimbulkan berbagai tantangan, terutama dalam interaksi sosial antar mahasiswa dan antara mahasiswa dengan dosen. Di UNESA, rendahnya kualitas interaksi antara mahasiswa dan dosen berdampak pada motivasi belajar dan kepuasan terhadap pembelajaran, sementara di UPJB, keterbatasan fasilitas dan aksesibilitas teknologi memperburuk kualitas pembelajaran daring. Penelitian ini berangkat dari asumsi bahwa keberhasilan akademik mahasiswa tidak hanya ditentukan oleh kemampuan kognitif, tetapi juga dipengaruhi oleh faktor afektif dan sosial, seperti interaksi, keterlibatan, dan kepuasan mahasiswa. Tujuan penelitian ini adalah untuk menganalisis pengaruh interaksi mahasiswa terhadap keberhasilan akademik, serta peran keterlibatan mahasiswa sebagai moderator dan kepuasan mahasiswa sebagai mediator. Penelitian ini menggunakan pendekatan kuantitatif melalui Structural Equation Modeling-Partial Least Squares (SEM-PLS) dan Random Forest Regressor. Sampel diperoleh dari mahasiswa Universitas Negeri Surabaya (UNESA) dan Universitas PGRI Jombang (UPJB). SEM-PLS digunakan untuk menguji hubungan linier antar variabel laten, sedangkan Random Forest Regressor digunakan untuk memodelkan hubungan non-linier dan mengevaluasi akurasi prediksi. Interaksi mahasiswa berpengaruh signifikan terhadap keberhasilan akademik, dengan koefisien jalur 0,302 di UNESA dan 0,117 di UPJB. Keterlibatan mahasiswa memperkuat hubungan tersebut sebagai moderator, dengan koefisien sebesar 0,645 (UNESA) dan 0,245 (UPJB). Kepuasan mahasiswa juga terbukti memediasi pengaruh interaksi terhadap keberhasilan akademik, dengan koefisien mediasi masing-masing sebesar 0,350 dan 0,355. Model Random Forest Regressor menunjukkan performa sangat baik dalam memprediksi keberhasilan akademik, dengan nilai R-Squared, MAE dan RMSE secara berturut turut di UNESA yaitu 0.9985, 0.0257, 0.1190 di UNESA dan 0.9984, 0.0267, 0.1228 di UPJB. Temuan ini menegaskan pentingnya membangun interaksi yang efektif serta meningkatkan keterlibatan dan kepuasan mahasiswa dalam pembelajaran daring.
Kata-kunci: Interaksi Mahasiswa, Kinerja Akademik, Kepuasan Mahasiswa, Keterikatan Mahasiswa, SEM-PLS, Random Forest Regressor

The development of online learning has brought significant changes in the higher education system, but it also poses various challenges, especially in social interaction between students and between students and lecturers. At UNESA, the low quality of interaction between students and lecturers has an impact on learning motivation and satisfaction with learning, while at UPJB, limited facilities and technological accessibility worsen the quality of online learning. This research departs from the assumption that students' academic success is not only determined by cognitive abilities, but also influenced by affective and social factors, such as student interaction, engagement, and satisfaction. The purpose of this study is to analyze the influence of student interaction on academic success, as well as the role of student involvement as a moderator and student satisfaction as a mediator. This study uses a quantitative approach through Structural Equation Modeling-Partial Least Squares (SEM-PLS) and Random Forest Regressor. Samples were obtained from students of the State University of Surabaya (UNESA) and PGRI Jombang University (UPJB). SEM-PLS is used to test the linear relationships between latent variables, while Random Forest Regressor is used to model non-linear relationships and evaluate the accuracy of predictions. Student interaction has a significant effect on academic success, with a path coefficient of 0.302 at UNESA and 0.117 at UPJB. Student involvement strengthened the relationship as a moderator, with coefficients of 0.645 (UNESA) and 0.245 (UPJB). Student satisfaction was also proven to mediate the influence of interaction on academic success, with a mediation coefficient of 0.350 and 0.355, respectively. The Random Forest Regressor model demonstrated excellent performance in predicting academic success, with R-Squared, MAE, and RMSE values of 0.9985, 0.0257, and 0.1190 respectively at UNESA, and 0.9984, 0.0267, and 0.1228 at UPJB. These findings emphasize the importance of fostering effective interactions and enhancing student engagement and satisfaction in online learning environments..
Keywords: Student Interaction, Academic Performance, Student Satisfaction, Student Engagement, SEM-PLS, Random Forest Regressor

Actions (login required)

View Item