Data Analytics for Effectiveness Evaluation of Islamic Higher Education using K-Means Algorithm

Abstract

The aim of this research is to utilize data analytics technology in evaluating the development of Indonesian national curriculum based on Indonesian National Qualification Framework, especially in universities. This research uses Exploratory Data Analysis (EDA) and several clusterization method, among others K-Means, K-Means++, MiniBatch K-Means, and MiniBatch K-Means++. The result of this research is not to measure the accuracy of clasterization result, but to discover the insight and interpretasion information from data collections that related with national curriculum in Indonesia. Based on the EDA and claterization methods with 30 variables of quetions and 67 students as respondent, MiniBatch K-Means with 2 cluster has the best pattern that reliable with highest Silhouette Coefficient value. However, on average K-Means++ has better interpretation than the others, with the average of Silhouette Coefficient value is highest. From that result, this research found that generally around 77,67% students can understand and feel the application of the Indonesian national curriculum well, but specifically only about 19.4% of students really understand and feel the impact of the curriculum very well. This is important to be evaluated by curriculum users in this case students and tertiary educational institution to improve the quality of academic services in the application of the Indonesian national qualification network.