Examiners Recommendation System at Proposal Seminar of Undergraduate Thesis by Using Content- based Filtering

Abstract

Undergraduate thesis is a student scientific activity which is accountable and also needs supervision and examination from lecturers to make sure it has a good quality. Therefore, supervisor and examiner should be the person that expert in a specific theme of undergraduate thesis. The purpose of this research is to build the examiners recommendation system on proposal seminar of undergraduate thesis. The method that applied is Content-based Filtering. Content-based filtering is applied because this research focuses in using the content of document. Undergraduate thesis report document is used as reference in this recommendation system. Undergraduate thesis report document is grouped based on the theme by using K- Means Clustering. The closeness of undergraduate thesis proposal is calculated from every centroid produced. The system will recommend which lecturers are in the cluster of nearest centroid. System testing is performed by measuring system performance using Ordered Analysis with Euclidean distance. The result of recommendation system has error value 0.385 which means the recommendation system has average level in the range of scoring 0-1. The accuracy of subset between recommendation result and actual data is 85%.