Pengenalan Perilaku Siswa dalam Proses Pembelajaran Menggunakan Progressive Evidence Distillation

Dwikara, Chudriana Naya (2024) Pengenalan Perilaku Siswa dalam Proses Pembelajaran Menggunakan Progressive Evidence Distillation. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002201031-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (9MB) | Request a copy

Abstract

Efektivitas proses belajar mengajar menjadi salah satu faktor utama dalam meningkatkan kualitas pendidikan. Lebih lanjut, efektivitasnya dapat diamati dari perilaku siswa selama proses belajar mengajar berlangsung apakah aktif atau pasif. Dengan mengenali perilaku siswa di kelas, dapat digunakan sebagai bahan pertimbangan guru dalam meningkatkan proses belajar mengajar. Oleh karena itu, penelitian Tugas Akhir ini bertujuan untuk mengenali perilaku siswa selama proses belajar mengajar berlangsung secara otomatis menggunakan model deep learning. Penelitian Tugas Akhir ini mengusulkan sebuah model integrasi Teacher-Student Curriculum Learning dengan Diagnostic Evidence Distillation (DED) bernama Progressive Evidence Distillation (PED) yang berusaha untuk mempelajari perilaku siswa dari fitur citra skeleton pergerakan siswa. Penelitian Tugas Akhir ini menguji model yang diusulkan pada dataset publik dari pengamatan siswa selama di kelas. Hasil eksperimen menunjukkan bahwa model PED mampu mengenali perilaku siswa dengan lebih baik dibandingkan model pembanding lainnya yang ditunjukkan melalui evaluasi metrik accuracy, precision, recall, dan F1 Score.

=============================================================
=============================================================

The effectiveness of learning processes is one of the main factors in improving the quality of education. We can know by observing how the students behave during the class and whether they are active. Further, teachers can improve the learning process by recognizing student behavior. Therefore, this study aims to recognize student behavior during the teaching and learning process automatically using a deep learning model. This study proposes an integration model of Teacher-Student Curriculum Learning with Diagnostic Evidence Distillation (DED) called Progressive Evidence Distillation (PED) that seeks to learn student behavior from the image features of student movement skeletons. This study tests the proposed model on a public dataset of student observations during class. Experimental results show that the proposed model is able to recognize student behavior better than the state-of-the-art models in terms of accuracy, precision, recall, and F1-Score.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Pengenalan Perilaku Siswa, Knowledge Distillation, Curriculum Learning, Student Behavior Recognition
Subjects:	Q Science > Q Science (General) > Q337.5 Pattern recognition systems Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Chudriana Naya Dwikara
Date Deposited:	07 Aug 2024 06:17
Last Modified:	07 Aug 2024 06:17
URI:	http://repository.its.ac.id/id/eprint/113733

Actions (login required)

View Item