Classification of Device Addiction to Students Using SAS-SV with K-Nearest Neighbor Algorithm Method

Abstract

A gadget is a small electronic device with a particular purpose, often thought of as an innovation of new goods. Not only to help facilitate human activities, but gadgets are also a part of the lifestyle for modern citizens. With this innovative feature, the gadget has attracted users more and more, or in other words, users have become more addicted to the gadget. This study aims to investigate how addictive gadgets are to students at the Department of Informatic Engineering, Khairun University, Ternate, Indonesia using K-Nearest Neighbor (KNN) Algorithm. In KNN, there is a Training dataset where one set of data contains the class's value and a predictor that will be used as one of the requirements for determining a suitable grade per the predictor. In contrast, the Testing dataset contains the new data that will be classified based on the model made and the accuracy of classification in the data collection process. Questionnaires were made using Google forms, then distributed through the internal groups of the Informatics Engineering department of  Khairun University. A total of 78 questionnaires were successfully collected. The results showed that the testing accuracy with k = 3 is 86% and k = 5 is 80%. This show that KNN algorithm can be applied to measure the level of addiction to students.