Forecasting Tingkat Kepuasan Siswa Terhadap Proses Pembelajaran Menggunakan Metode K-Nearest Neighbor

Abstract

Education is a crucial aspect in the formation of society and individual development. Student satisfaction is an important indicator that reflects the effectiveness of the learning process. In this context, this research proposes applying the K-Nearest Neighbor (KNN) method as a predictive tool to identify factors influencing student satisfaction. Historical data on student satisfaction is collected and analyzed to build a prediction model using KNN. This research aims to increase the effectiveness of the learning process by understanding the factors contributing to student satisfaction. Through personalizing learning experiences, identifying causes of dissatisfaction, and developing innovative strategies, KNN predictions can provide deep insights into educational institutions. It is hoped that the results of these predictions can be used to increase student retention, efficiency in academic management, and transparency in the educational environment. By integrating artificial intelligence into evaluating student satisfaction, this research contributes to developing more adaptive and responsive educational strategies. In conclusion, predicting student satisfaction using KNN is an essential basis for creating a learning environment that has a positive and sustainable impact on student development in the modern educational era. This data was collected by distributing questionnaires to students with a sample size of 160 data. So it is known that there are 112 training data and 48 testing data. Then, from applying the K-Nearest Neighbor method, the value of K=12 is known. So, the test results using the Python programming language with a data division of 70%:30% produce an accuracy value of 80%, a precision value of 79%, a recall of 100%, and an F1-Score of 88%.