Construction Research on Quality Feature Dictionary of Medical Services Based on Tongyici Cilin and Word2vec Model

As the foundation work for natural language processing tasks such as sentiment analysis and text classification, the quality and accuracy of feature dictionaries directly affect the performance of models and the precision of results. Given the particularity of the medical services field, there is still a shortage of high-quality feature dictionaries in the medical services domain. In this paper, we utilize patients' medical services review corpus to select high-frequency seed words and classify them using the TF -IDF algorithm. Then, we expand the feature words by calculating the similarity between seed words using Tongyici Cilin (a Chinese synonym dictionary) and Word2vec, aiming to find a more precise and comprehensive feature extraction method, and construct a feature dictionary for medical services quality. This method is concise and efficient, and can effectively improve the sentiment analysis and text classification effects for medical texts, playing a key role in the field of medical services.