Analysis of Poor Population in DKI Jakarta Regions using Fuzzy C-Means and K-Means Algorithms
Syafirina Arsyidin Sidqi, Rafian Tri Akso, Imam Bayu Nurdianto, Alia Novendri, Imam Tahyudin, Siti Alvi Sholikhatin · 2022
The government of DKI Jakarta have been struggling to lower the number of poverties in its region. The first step that the government need to do is analyzing the number of poverty itself. The information is important to identify the effective way to solve the problem based on the priority of the region that have most poverty number. By now, the government has formulated the data collection by weighting the 15 indicators into 3 groups. The large number of data and indicators may cause difficulties in its manual implementation, make the process become ineffective and less objective. Therefore, automation is needed in the process of clustering poverty data. This study aims to analyze the performance of the FCM algorithm (Fuzzy C-Means) and K-Means which is implemented on poverty data in Central and categorize it into 3 clusters. There are several steps that must be carried out before clustering step: 1) preprocessing, namely data cleaning and data transformation, 2) clustering is carried out using the two algorithms (FCM and K-Means). The calculation results are used to compare between two algorithms. Based on the comparison result, it is obtained that FCM is better than K-Means algorithm, which has an accuracy rate of only 83.33%, which better than the K-Means algorithm of 50%.