Chinese News Hot Subtopic Discovery and Recommendation Method Based on Key Phrase and the LDA Model
Bin Ge, Chunhui He, Shengze Hu, Cheng Guo · DEStech Transactions on Engineering and Technology Research · 2018
The discovery and recommendation of Chinese news hot subtopic is a growing research area, and currently technology in this area is not yet mature. This research, inspired by traditional LDA model, uses the key phrase as feature to construct a “bag-of-phrase”. On this basis, we propose the Chinese news hot subtopic discovery and recommendation method based on key phrase and the LDA model. Hot subtopic discovery and recommendation is the mainly innovation in this paper. During the cluster of hot subtopics, we select using the Longest Common Sequence (LCS) value as the similarity distance. To evaluate the proposed method, we used a mixed Chinese news dataset to experiment, and adopting: (1) time consumption of the training model, and (2) Perplexity value, and (3) quality of the discovery hot subtopic, three index to evaluate the performance of the method. The experimental results show that the proposed method can accurately discover the news hot subtopics, and also efficiently recommend relational hot subtopics in various fields.