Pengembangan Metode Hibrida Transformasi Data Dan Meta-Algoritma: Terapan Untuk Identifikasi Level Kognitif Siswa MIPA

Yamasari, Yuni (2020) Pengembangan Metode Hibrida Transformasi Data Dan Meta-Algoritma: Terapan Untuk Identifikasi Level Kognitif Siswa MIPA. Doctoral thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 07111660010011_Dissertation.pdf]
Preview
Text
07111660010011_Dissertation.pdf

Download (9MB) | Preview

Abstract

Kemajuan Teknologi Informasi dan Komunikasi (TIK) pada area pendidikan memungkinkan untuk merekam semakin banyak karakteristik siswa. Sebenarnya, hal ini sangat menguntungkan karena semakin banyak karakteristik siswa maka semakin akurat guru dalam melakukan proses penentuan prestasi belajar siswa, khususnya ranah kognitif siswa. Namun, disisi lain, keadaan ini mengakibatkan guru menemui kesulitan dalam proses penentuan ini karena semakin banyak karakteristik siswa yang harus dipertimbangkan. Kondisi ini merangsang penelitian tentang pemodelan prestasi belajar siswa, khususnya, dalam rangka penentuan ranah kognitif siswa. Namun, solusi dan strategi yang ditawarkan untuk penentuan ranah kognitif masih sangat terbatas. Bahkan, solusi yang ada masih mempunyai kinerja yang rendah sehingga kesalahan identikasi masih tinggi. Hal ini disebabkan eksplorasi terhadap metode-metode pada tiap tahap pembangunan system identifikasi level kognitif ini belum banyak dilakukan. Untuk itu, penelitian disertasi ini memfokuskan pada eksploitasi metode-metode pada masing-masing tahap agar sistem yang dihasilkan dapat mereduksi kesalahan identifikasi sebanyak mungkin. Upaya-upaya yang dilakukan ini akhirnya dapat menghadirkan metode hibrida transformasi data dan meta-algoritma LogitBoost untuk membangun sistem identifikasi level kognitif siswa MIPA secara lebih optimal. Hasil eksperimen menunjukkan bahwa sistem yang dibangun oleh metode hibrida yang diusulkan ini mempunyai keunggulan dibandingkan hanya menggunakan metode-metode tree (Decision Stump, REP Tree, Random Tree) saja. Hal ini diindikasikan dengan kenaikkan akurasi tertinggi hingga 21,062% dari 76.903% ke 97.965% dan kemampuan reduksi terbanyak dari kesalahan identifikasi level kognitif siswa MIPA sebanyak 24 siswa dari 26 ke 2 siswa.
===============================================================================================================================
The advancement of Information and Communication Technology (ICT) in the education area makes it possible to record a growing number of student characteristics. Actually, this is very beneficial because the more characteristics of students the more accurate the teacher are in the process of determining student achievement, especially the cognitive domain of students. However, on the other hand, this situation causes the teacher to encounter difficulties in this determination process because more and more student characteristics must be considered. This condition stimulates research on modeling student achievement, in particular, in the context of determining the cognitive domain of students. However, the solutions and strategies offered for determining cognitive domains are still very limited. In fact, existing solutions still have low performance so that the identification error is still high. This is due to the exploration of methods at each stage for developing of the identification system on the cognitive level that has not been done much. For this reason, this dissertation research focuses on the exploitation of methods at each stage so that the system resulted can reduce as many misidentifications as possible. Finally, the efforts worked can present the hybrid of data transformation method and the LogitBoost as meta-algorithm to build a system identifying the cognitive level of MIPA students. The experimental results show that the system built by the proposed hybrid method has advantages over using only tree methods (Decision Stump, REP Tree, and Random Tree) alone. This is indicated by the highest accuracy increase up to 21,062% from 76,903% to 97,965% and the highest reduction ability from the misidentification of the cognitive level of MIPA students by 24 students from 26 to 2 students.

Item Type: Thesis (Doctoral)
Additional Information: RDE 006.312 Yam p-1
Uncontrolled Keywords: LogitBoost, cognitive domain, feature selection, feature extraction, feature discretization
Subjects: L Education > L Education (General)
L Education > LC Special aspects of education > LC5800 Distance education.
T Technology > T Technology (General) > T57.5 Data Processing
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20001-(S3) PhD Thesis
Depositing User: YUNI YAMASARI
Date Deposited: 14 Mar 2025 03:56
Last Modified: 14 Mar 2025 03:56
URI: http://repository.its.ac.id/id/eprint/75243

Actions (login required)

View Item View Item