<![CDATA[The rapid growth of scientific publications has made it increasingly difficult for researchers to efficiently identify relevant literature from multiple scholarly sources. Integrating heterogeneous scientific databases with an effective recommendation mechanism can improve both the relevance and efficiency of literature retrieval. This study aims to design and implement a scientific journal aggregator system that integrates multiple scholarly databases through Application Programming Interface (API) technology and incorporates a text mining-based recommendation mechanism to improve the relevance of search results. The proposed system integrates article metadata from OpenAlex, Semantic Scholar, and the Directory of Open Access Journals (DOAJ). Text preprocessing, including case folding, tokenization, and stopword removal, is applied before calculating Term Frequency–Inverse Document Frequency (TF-IDF) weights and Cosine Similarity to measure the relevance between user queries and scientific articles. The proposed system was evaluated using multiple search keywords, including Artificial Intelligence, Machine Learning, and Technology, to assess its recommendation performance across different research topics. Experimental results demonstrate that the system consistently achieved high recommendation performance, with accuracy ranging from 95.32% to 97.95%, precision ranging from 95.24% to 97.53%, recall consistently reaching 100.00%, and F1-score ranging from 97.56% to 98.75%. These findings indicate that integrating multiple APIs with TF-IDF and Cosine Similarity provides an effective approach for improving the relevance, coverage, and efficiency of scientific literature retrieval across different search topics.]]>
