Open AccessPrediksi Hasil Belajar Siswa Secara Daring pada Masa Pandemi COVID-19 Menggunakan Metode C4.5
Author(s) -
Yetti Fitriani,
Sarjon Defit,
Gunadi Widi Nurcahyo
Publication year - 2021
Publication title -
jurnal sistim informasi dan teknologi
Language(s) - English
Resource type - Journals
ISSN - 2686-3154
DOI - 10.37034/jsisfotek.v3i3.54
Subject(s) - covid-19 , attendance , test (biology) , pandemic , computer science , sample (material) , mathematics education , machine learning , artificial intelligence , psychology , medicine , paleontology , chemistry , disease , pathology , chromatography , infectious disease (medical specialty) , economics , biology , economic growth
Student learning in schools has changed since the Covid-19 pandemic. Student learning in normal conditions is carried out face-to-face and turns into online or online learning. The research was conducted to predict student learning outcomes during the COVID-19 pandemic so that the results of this study can be used as a reference in policymaking in schools. The C4.5 method was used in the study to classify the data for class XII of the Multimedia Department at SMKN 2 Padang Panjang and the classification results could predict student learning outcomes during the pandemic. Processed student value data were taken from 1 (one) subject as the research data sample. Analysis of the value of student learning outcomes using the C4.5 Method to obtain new knowledge from student learning outcomes data carried out during the COVID-19 pandemic. The data analyzed consisted of attributes of attendance, assignments, daily tests, and test scores which influenced the decision criteria for student learning outcomes in online learning. The learning outcome decision criteria consist of "Satisfactory" and "Not Satisfactory" which refer to the Minimum Completion Criteria. Tests conducted on the training data of learning outcomes show that the value of the Daily Test is the most influential attribute in decision making. Implementation of the results using the RapidMiner Studio 9.2.0 software and produces an accuracy of 83.33% of the test data testing with the rules of data analysis training results. The results of the C4.5 classification testing method in this study can be used to predict student learning outcomes. The test results with an accuracy of 83.33% can be recommended to help schools in making policies