Semantic similarity of Indonesian sentences using natural language processing and cosine similarity

Natural language processing is a commonly used method for processing text input. The text is widely used in many ways, such as determining hoaxes, concluding sentences, sentiment analysis, and measuring similarity. Many other studies in English have carried out measurements of similarity between two sentences. However, there are still few studies that use Indonesian sentences. This study proposes an approach to assess the similarity between two Indonesian sentences. The motivation of this research is to prove that the proposed approach can be applied. Similarity assessment uses a combination of natural language processing and cosine similarity. Natural language processing begins with POS tagging, tokenization, stemming, and similarity measurement between words using WordNet and Wu Palmer. The results showed that the accuracy value is 86,67%. Therefore, this study can be used as a preliminary process in natural language processing with more complex objectives.