Topic discovery method based on topic model combined with hierarchical clustering
An Wang, Junjie Zhang · 2020 IEEE 5th Information Technology and Mechatronics Engineering Conference (ITOEC) · 2020
Topic discovery is increasingly used today in the rapid development of the Internet, especially in the field of Weibo. However, due to the short content of Weibo, the traditional LDA topic model has some effects on the topic discovery of Weibo. An LDA model combined with an improved BIRCH hierarchical clustering algorithm was used for topic discovery, and short text extensions were made for microblogs with short text content. Compared with the combined effect of K-means and BIRCH clustering algorithms, the results of topic discovery using BC-BIRCH clustering algorithm are better.