Data Cleaning and Prototyping Using K-Means to Enhance Classification Accuracy

This research is to propose a preprocessing method in a data classification process to improve the accuracy of the classification process. The preprocessing method to be used basically aims to clean up the data by forming prototype data generated from the clustering process. It is expected that the proposed preprocessing methods can be used in a variety of data forms and classification algorithms. The difference of this study is the preprocessing stage which consist of 2 processes, which are, the process that eliminate missing value and make data prototyping using clustering approach. After performing the preprocessing steps, the complete and evaluated or clustered data as prototype data, where the cluster will refer to certain classes would be obtained. After the data prototype was obtained, the main steps would be started which consist of the classification and evaluation process. The cycle was generate the classification accuracy results. Based on the comparison of classification accuracy without using and using preprocessing, by using the method proposed in this study could increase approximately 34% accuracy results or reduce the error 75.55%.