Fuzzy C-Means Clustering Untuk Identifikasi Keaktifan User Berdasarkan Log Data Pada Portal UAD.ac.id
Oleh Farid Bagas Sasongko dan Dewi Pramudi Ismi
Abstract:
This study aims to group user portal.uad.ac.id students into 3 groups seen from their activeness in accessing the academic portal. This system classifies users who are included Frequently, moderately, and lacking in opening portal sites. This research was made to classify activeness user, so it is hoped that students will be more critical in opening the academic web portal for smooth running academic process on campus. The method used in this research is Fuzzy C-Means. System development stages starting with analysis, design, implementation, and testing. Flow system by cleaning data (cleaning), data selection (data selection) takes data attributes that will be used, data transformation (data transformation) transforms data from text to numbers to take values and the Clustering process with the Method Fuzzy C-Means. The clustering process runs by inputting the number of clusters, epsilon, rank and maximum iteration. The research conducted resulted in a web-based user grouping system that could determine which students are active, sufficient and lacking in accessing the portal. Alpha Test test results It was found that the system for grouping users based on log data was functionally feasible with the results reached 90% of respondents accepted. While testing with the Dunn Index below is based on the results clustering 2/3 is worth 1 or more than 3 clusters, so the results of clustering are tight and well separated (well separated). These results are expected to be able to solve the problem of information that has not reached the user.
Key Words: fuzzy c-means, clustering, web log data, web usage mining
