<![CDATA[The rapid growth of scientific publications in the field of computer science has created a need to understand the distribution and trends of emerging research topics. This study aims to identify and analyze dominant topics in computer science literature using a text mining approach based on Term Frequency–Inverse Document Frequency (TF-IDF) vectorization and the K-Means clustering algorithm. A total of 1,222 publication titles from Semantic Scholar (2020–2025) were processed through language normalization, text preprocessing, TF-IDF feature extraction, optimal cluster determination, and cluster quality evaluation using Silhouette Score and Davies-Bouldin Index (DBI). The results reveal that topics such as cybersecurity, artificial intelligence, and machine learning are the most prevalent. While some clusters show good internal cohesion, the overall evaluation yielded a Silhouette Score of 0.0585 and a DBI of 4.387, indicating overlapping topics and limited cluster separation. These findings suggest that although the TF-IDF and K-Means approach can highlight general topic trends, it has limitations in capturing semantic context. Future research is encouraged to explore more contextual representation and clustering techniques to improve topic analysis quality.]]>
Copyrights © 2025
