EXPERIMENTAL COMPARISON OF MACHINE LEARNING ALGORITHM PERFORMANCE FOR OPTIMIZING ELECTIVE SUBJECTS - Application in Merdeka Curriculum Phase F

Aug 5, 2026

Upload by Liam

In Phase F of the Merdeka Curriculum, choosing electives is a critical task for grades 11–12, where students must select four out of twelve options aligned with strengths, interests, future study plans, and career goals over two years. Machine learning is used to improve the effectiveness and efficiency of elective selection, leveraging 10th-grade report card data and results from interest, aptitude, further study, and career-choice tests plus prior manual selections. Cross-validation supports generalization for the small dataset. Experiments compare five small-dataset algorithms—Naive Bayes, Decision Tree, Random Forest, SVM, and KNN—using accuracy, recall, and confusion matrix. Results show SVM delivers the best performance (highest accuracy 57.3770%, recall 0.574, and true positive 0.574).

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Abstract

Introduction

Background

Merdeka Curriculum Phase F elective selection context

EXPERIMENTAL COMPARISON OF MACHINE LEARNING ALGORITHM PERFORMANCE FOR OPTIMIZING ELECTIVE SUBJECTSELECTION IN PHASE F OF THE MERDEKA CURRICULUM Dedy Mulyadi*1 1Informatics, Universitas Pradita, Indonesia [Email:](Email:1dedy.mulyadi@student.pradita.ac.id)[1](Email:1dedy.mulyadi@student.pradita.ac.id)[dedy.mulyadi@student.pradita.ac.id](Email:1dedy.mulyadi@student.pradita.ac.id) (Article received: April 15, 2024; Revision: May 06, 2024; published: August 01, 2024) Abstract In phase f of the Merdeka Curriculum, electives are an important element at the senior high school level. Students are faced with the challenge of choosing four out of twelve elective subjects that are relevant to their talents, interests, further study plans and career goals over a two-year study period. Applying machine learning with the right algorithm is a solution for the effectiveness and efficiency of elective selection. The dataset used comes from the 10th grade report card data, the results of the interest, aptitude, further study, and career choice tests, and the manual selection of electives chosen by students in the previous year. The use of a small data set requires a crossvalidation method to improve the generalizability of the model and to optimize the data set, thereby increasing the validity of the results. The test will be conducted using an application that tests five machine learning algorithm models suitable for small datasets, namely Naive Bayes, Decision Tree, Random Forest, Support Vector Machine, and k-Nearest Neighbors. The test focuses on comparing the performance of the five algorithms based on the best accuracy, recall, and confusion matrix and the results obtained Support Vector Machine (SVM) algorithm has the best performance results by achieving the highest accuracy of 57.3770%, the highest recall of 0.574, and the highest true positive (TP) of 0.574. The Support Vector Machine (SVM) algorithm will be a recommendation for further research, namely the development of machine learning for the selection off-stage elective subjects atAtisa Dipamkara senior high school, to provide relevant guidance to students in making decisions regarding the selection of elective subjects more accurately and according to their respective characteristics. Keywords: cross validation, decision tree, KNN, machine learning, merdeka curriculum phase f, naive bayes, random forest, small dataset, SVM. 1. INTRODUCTION 1.1. Background In the world of education, the fundamental for the implementation of an education system is the curriculum which is a set of plans and arrangements regarding the objectives, content, and learning materials as well as the methods used as guidelines for organizing learning activities to achieve curriculum objectives. In Indonesia itself, the curriculum has undergone several changes, ranging from the 1947 curriculum, 1994 curriculum, 2006 curriculum, 2013 curriculum, to the Merdeka Curriculum used in most schools in Indonesia today. The Merdeka Curriculum is a curriculum with diverse intra curricular learning where content will be optimized so that learners have enough time to explore concepts and strengthen competencies [1] . In the Merdeka Curriculum there are have 6 phases, namely: Phase A for grades 1-2 elementary school, Phase B for grades 3-4 elementary school, Phase C for grades 5-6 elementary school, Phase D for grades 7-9 junior high school, Phase E for grades 10 senior or vocational high school and Phase F for grades 11- 12 senior or vocational high school. In phase f of the Merdeka Curriculum for grades 11 and 12 of senior high school there are elective subject groups that each student can choose to pursue for 2 years of study. As stated in the Decree of the Minister of Education, Culture, Research and Technology of the Republic of Indonesia number 262/M/2022 concerning guidelines for curriculum implementation in the framework of learning recovery on pages 21-23, the curriculum structure for ph

FAQs

What problem does the study address in Merdeka Curriculum Phase F?

The study addresses how students in grades 11–12 select four electives from twelve options in a way that matches their talents, interests, further study plans, and career goals over two years.

Which machine learning algorithms are compared in the experiments?

The experiments compare Naive Bayes, Decision Tree, Random Forest, Support Vector Machine (SVM), and k-Nearest Neighbors (KNN).

Why is cross-validation used in this research?

Cross-validation is used because the dataset is small, improving the model’s generalizability and optimizing the dataset to increase the validity of the results.

Aug 5, 2026

Upload by Liam