<![CDATA[Social media platforms such as Twitter present a wide range of emotional expressions from users in short and informal texts, which pose significant challenges for automated analysis. This study develops a search system for Indonesian-language tweets that identifies user emotions based on semantic similarity to a given text query. The method employs Term Frequency-Inverse Document Frequency (TF-IDF) for feature weighting and Cosine Similarity to measure textual similarity. Preprocessing stages including normalization, tokenization, stopword removal, and stemming are applied to enhance text representation accuracy. The system is tested using emotion-based queries and returns relevant tweets with high semantic match scores. Experimental results show that 50% of the top retrieved tweets match the expected emotional context. This approach proves effective in detecting emotions in short texts and offers potential for further development in sentiment-driven opinion analysis and emotion-aware recommendation systems]]>
Copyrights © 2025
