TF-IDF Implementation for Similarity Checker on The Final Project Title

Abstract

Students of the Software Engineering Technology Study Program, Department of Information Technology, Politeknik Negeri Padang, is required to compile a final project to complete their study period. In the implementation of the final project, several parties have involved such as the KBK team whose job is to check whether the proposed title is appropriate or not. The main issue undertaken by KBK is whether the title submitted has been used or not. The method used by KBK in checking the availability of titles was by looking at the titles of the final projects that have been submitted by previous students. The examination process carried out by the KBK, it took a long time. By utilizing the Cosine Similarity algorithm and TF-IDF, it is expected that it will make it easier for KBK to check the availability of final project titles. Cosine similarity is a method used to calculate the degree of similarity between 2 or more documents. While the TF-IDF Algorithm is a method used to weight a word in a document. The object of testing in this study was the title of the student’s final project. The process of calculating the level of similarity of documents started from the preprocessing stage, then proceeds with weighting using TF-IDF and calculating the level of similarity used the Cosine Similarity algorithm. The final result found that system could calculate the degree of similarity of the title of the student’s final project. From the results of testing the process of calculating the degree of similarity of titles using the cosine similarity algorithm can be undertaken quickly.