Thesis is a compulsory subject that must be completed by students to obtain undergraduate status (S1). Students' need for thesis information is increasing, so labeling topic trends is expected to help students find out which topics are trending in the previous year without having to read the entire thesis in the library. In this study, thesis data collection was based on data from the previous 5 years, namely 2015 - 2020 at the Department of Informatics, Universitas Malikussaleh (Unimal). Thesis topic trends are classified into 5 categories, namely Data Mining, Artificial Intelligence, Image Processing, DSS, and GIS. Classification of thesis topics based on title and abstract. The programming language used is PHP and MySQL database. The methods used in the automation of thesis topic trend determination are text mining and the K-Means Clustering algorithm. The priority of this research is to produce an application that can overcome student obstacles in determining thesis topics. This application is expected to make it easier for students to find out trends in thesis topics in the previous year. In this study, the percentage accuracy of the thesis topic trend using the K-means clustering algorithm is 84% of the 70 test data
Copyrights © 2022