ANALISIS PEMETAAN CRIME PATTERN MENGGUNAKAN CLUSTERING K-MEANS

Main Article Content

Helga Eva Julia
Bagus Mulyawan
Dedi Trisnawarman

Abstract

Data mining is one of the way to obtain important information from huge database. People can obtain useful information and decide future actions. Generally, data mining process consist of 6 phase, i.e. business understanding, data understanding, data preparation, modeling, evaluation, and deployment. K-Means clustering was used to divide data into groups for this study. This study aims to define the best K-Means clustering model used for Indonesian criminal database mapping based on number of cluster. Experiment was conducted with various number of records, i.e. 100, 1.000, 20.000, 50.000, 80.000, and 126.000. Experiment was also conducted with both 2 cluster and 3 cluster. Clustering result was compared with clustering result from WEKA machine learning system to determine if K-Means clustering has been done properly. Test result showed that    system K-Means clustering with 2 cluster give better results by 65.33 % accuracy with error rate 34.77, when system K-Means clustering with 3 cluster give results by 53.5% with error rate 46.5%..

 

Key words

Clustering, Crime Pattern, Data mining, K-Means, Mapping

Article Details

Section
Articles