Classification Of Eligibility to Receive Social Assistance Using The K-Nearest Neighbor Method

Abstract

Data classification is the process of grouping data based on attributes (Congregation Employment, Congregation Dependents, Congregation Home Status, and Congregation Income). The problem currently occurring is that data collection on congregational social assistance recipients is often not on target, so that if social assistance enters the church, it is given to congregations who are actually less well off, but it is transferred to congregations who are well off, giving rise to confusion between one congregation and another. In this research the author used the K-Nearest Neighbor method or what is usually called KNN and measured algorithm performance using a confusion matrix to calculate accuracy, precision and recall. This researcher used 50 data that had been input via Google Form and then filled in the congregation from 50 data divided into 35 training data and 15 testing data. After the data has been input it will go through several stages, the first step is initialization where in the process of this initialization stage it changes the category value, the second stage is the process of dividing the value by the largest value in the attribute and the third stage is calculating the distance to then calculate the confusion matrix to determine accuracy, precision and recall. This research produces an application that can automatically determine which congregations are and are not worthy of receiving social assistance. From trials of 15 testing data, accuracy was 88.89%, precision 100% and recall 75%