Prediction Of Election Participant With Malang City Demographic Data Using The K-Nn Algorithm

Abstract

Election (General Election) is a step to choose and determine the figure of a leader. In general election activities, the higher the level of political participation indicates that the people understand the importance of democracy. On the other hand, if the level of participation is low, the people are less concerned about state problems. From the 2019 election activities in Malang City, the next step is connecting with demographic data sourced from the Central Statistics Agency (BPS). The demographic data includes aspects of Energy, Geographic, Education, Health, Population, Economy, Communication, Transportation, and Expedition, which are then integrated with election data. In the 2019 Presidential Election, the number of DPT (Permanent Voters List) was 623,185, while the number of citizens who exercised their right to vote was only 488,587. This study will look at the relationship of demographic data to public participation in implementing elections. Using the classification method for prediction with high and low label classes on the form of community participation at the polling station (TPS) level. In the preprocessing stage, the dataset model is determined by testing three types of normalization methods, then implemented in the K-Nearest Neighbor (K-NN) algorithm. From the test results, the highest level of accuracy obtained in predicting voter participation is 61.83%, and the F-1 score is 61.46%, with the Min Max normalization model occupying the best results.