<![CDATA[The process of selecting a thesis advisor that aligns the thesis topic with the advisor's area of expertise will greatly assist in the thesis preparation process for students. However, in reality, the process of determining the thesis advisor is often not based on the advisor's area of expertise. The process of determining the thesis advisor is often done manually and inefficiently, which can lead to a mismatch between the advisor's area of expertise and the student's thesis topic. From this problem, this research developed a recommendation model for selecting thesis supervisors using the Cosine Similarity method, utilizing thesis title data and the list of lecturers' areas of expertise. The process begins with preprocessing the text data of thesis titles, then calculating word weights using TF-IDF and measuring similarity to finally match the students' thesis titles with the lecturers' areas of expertise, which are calculated using the Cosine Similarity algorithm. From the results of testing the process of matching research topics from student thesis titles with the expertise and research of lecturers, it was produced accurately, as indicated by a precision value of 81% and a recall of 90%. This result also shows that the relevance level of the recommendations is very high. The F1 score obtained, which is 85%, indicates that the modeling has a good balance between precision and sensitivity (recall).]]>
Copyrights © 2025
