KLUSTERISASI LOG AKTIFITAS PADA SISTEM KEAMANAN FILE DENGAN PENDEKATAN DATA MINING

Abstract

The file system is an area or resource that can be considered critical in addition to having high sensitivity. The file system is related to user management and is monitored through the activity history log in it. Users can interact fully in the file system equipped with their permissions and rights. Activities on the file system will become more complex and related to security issues in the file system. Thus, the protection of the file system is a major concern. In this study, an analysis of user activity history logs in the file system will be carried out using a data mining approach. The method applied in the analysis uses the clustering method. The purpose of clustering is to group or segment users regarding activities on the file system. The K-Means technique is applied to the clustering method and the performance results are presented by identifying the findings of 5 clusters in the activity log. In an effort to ensure that the model or clustering method using the K-Means technique is accurate, the Davies Boulding and F-Measure methods are used to ensure the quality of accuracy in the clustering method. Therefore, the clustering method with the K-Means technique is a method that is considered quite good in grouping user data related to activities on the file system.