Dynasty novels are the golden period of Chinese literary development, in which dress metaphors, as an important means of literary expression, contain rich cultural connotation and social significance. This study focuses on dress metaphors in Ming Dynasty novel texts, and constructs a semantic network of dress metaphors through the improved TF-IDF algorithm, latent semantic analysis method, and semantic network analysis method to mine the hidden information in Ming Dynasty novel texts. The study constructed a corpus containing 4580 sentences, including 2677 non-metaphorical sentences and 1903 metaphorical sentences. The quantitative calculation of the dress words by LSA method reveals that the dress words with the highest correlation in the original and metaphorical meanings are the tonic and the wuzi cap, with the correlation of 0.114241 and 0.124387, respectively. The analysis of Water Margin shows that the red headscarf, as a symbol of the resistance against oppression of the Liangshan heroes, has the highest frequency of appearing throughout the book, with a frequency of 218 times; and the wuzi cap, which symbolizes the bureaucratic system, and tonic appear 188 and 182 times respectively, reflecting the twists and turns of the characters’ fates in the work. The semantic network analysis further reveals the intrinsic connection between dress metaphors and the novel’s themes, and shows how dress words carry social and cultural information and changes in characters’ fates by means of metaphors. This study provides new computational methods and theoretical perspectives for analyzing the texts of Ming Dynasty novels, and helps to deepen the understanding of metaphorical expressions in Ming Dynasty literature.
