Modern Chinese literature contains rich thematic connotations. This study proposes a method for extracting and analyzing themes of modern Chinese literature texts based on vector space model. Firstly, the text is preprocessed, including data cleaning, word splitting and deactivation word removal; then the text is transformed into multi-dimensional vector representation by using vector space model and the text feature weights are calculated by TF-IDF; finally, a two-stage clustering strategy is designed, in which the number of class clusters and centers are estimated by Canopy algorithm, and then fine classification is performed by K-means algorithm. The experimental results show that when the number of topics is set to 7, the model perplexity is the lowest at 6.646, the clustering precision rate reaches 0.81, the recall rate is 0.796, and the F-measure value is 0.802, which is obviously better than other settings of the number of topics. By analyzing 31,226 data of modern Chinese literature, seven major themes are successfully extracted: criticism of nationalism, oppression of feudal rites, enlightenment and salvation, cultural conflict between urban and rural areas, dilemma of women’s awakening, writing of war sufferings, and uncertainty of intellectuals. The study shows that the vector space model combined with the optimized K-means algorithm can effectively identify the thematic features in modern Chinese literature and provide data support for literary research.
