Forecasting: Analyze Online and Offline Learning Mode with Machine Learning Algorithms

Abstract

Since the pandemic occurred, in March 2020, learning activities have changed from an offline to an online learning mode. This is the first time, such a huge change has occurred, simultaneously in the entire hemisphere. This learning mode opens a new discourse regarding the impact on the learning mode and educational evaluation results. The author aims to compare the results of the educational evaluation of the online learning mode during the pandemic with offline learning mode, so that differences will be known, as well as can be used to predict student learning outcomes, in order to obtain an overview of the effectiveness and efficiency of a learning mode. Data collection is carried out as an initial step in data processing, based on the final results of student learning, in certain courses taken every semester starting in 2017-2022. The data consists of 6 indicators, namely CI1-CI4, grades, and letter grades. The result of this study is the prediction of a more effective learning mode used, as decision support carried out by the forecasting method, comparing the Naïve Bayes and Decision Tree algorithm in getting the best accuracy value, by analyzing the learning mode offline to online.