A short text topic discovery method for social network

Short text theme discovery is the discovery of hot topic from short text data in mass. As the micro-blog social network has distinct characteristics of the network language, new words emerge in an endless stream. This paper presents an improved method for short text theme found, First, based on HMM model discovered new words to the text, new words are added to the user dictionary, and then we use discovery results of new words to build LDA model, finally, get the document clustering-topic distribution. The experimental results show that this method can effectively enhance the comprehensiveness and accuracy of the topic discovery and is more suitable for theme mining under social network environment.