ANALYSIS AND IMPLEMENTATION OF ALGORITHM CLUSTERING AFFINITY PROPAGATION AND K-MEANS AT DATA STUDENT BASED ON GPA AND DURATION OF BACHELOR-THESIS COMPLETION
Achmad Benny Mutiara, Asep Juarna · 2012
Effectiveness and accurate results from an algorith m has always been a basic reference for every step taken in the use and utilization of algorithm, which is e xpected to achieve optimal results both in quality and quantity. In order to realize the level of accuracy and effectiveness from the program, it would requi re an algorithm that can minimize error and faster in dat a processing rate compared with existing algorithm In this paper, we have compared two algorithms, namely Affinity Propagation and K-Means, at data student based on GPA and Duration of Bachelor-Thesis Completion. The results show that Affinity propagation gives the result of data cluster more accurate and effective than K-Means, it can be seen from the tes ting table which showing that the value of affinity prop agation exemplar has not changed at all after five trials. While K-Means, gives values of its centroid are dif ferent after five trials. And at the data students itself, it show that there is a relationship between GPA and D uration of Bachelor-Thesis completion in Gunadarma University students, it can be seen from the result s of data clustering, that is for student who have GPA above 3 to 4 have a tendency to finish their Bachel or-Thesis faster, which is less than 1 until 2 seme sters. While other students who have GPA less than 3 have a longer time to finish their Bachelor-Thesis, with in a period of 2 until more than 4 semesters.