<![CDATA[Background: A thesis is a course required for completing a Bachelor's degree (S-1). Thesis documents are typically collected in a library file. This data will be more useful if subjected to in-depth analysis. One such analysis is identifying trends in student thesis topics using clustering techniques. Objective: To analyze and group thesis titles from the years 2012 to 2017 using text mining techniques, with the aim of identifying topic trends through clustering with the K-Harmonic Means method. Methods: The method used in this study is K-Harmonic Means clustering. The stages carried out include tokenization, filtering, stemming, TF-IDF, grouping using K-Harmonic Means, and testing using purity. Results: The result of this research is an application that can process thesis titles into trend groups of thesis titles. From the test conducted using purity obtained a value of 0.63 which means the K-Harmonic method is quite good in grouping. The results of the analysis show that the topic of Multimedia and soft computing became a trend for 3 years, namely 2012, 2013 and 2014, while the topic of mobile applications and web programming became a trend in 2013 and 2015. Conclusion: The results of grouping using the K-Harmonic Means method show a sufficient accuracy value of 0.63. This proves that the K-Harmonic Means method is quite suitable for carrying out the process of grouping text-based data.]]>
Copyrights © 2025
