Research Based on the HowNet Semantic Expansion Micro-Blog Hot Topic Detection System

Abstract:

In this paper, we propose an improved topic detection method based on the HowNet semantic expansion. Using the HowNet to extent the VSM of micro-blog content, solve the problem that the similarity calculation is not accurate because of the length of the micro-blog text is too small and the gap of user comments number is too large through the two-step above. The similarity calculated accurately, would there be a good clustering effect. Put forward a more complete topic heat assessment methods that take into account the number of fans of the Micro-Bo users, the number of forwarding Micro-Bo text, comments, and time factors.

• Periodical: Advanced Materials Research (Volumes 765-767)
• Pages: 1502-1506
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.765-767.1502
• Online since: September 2013
• Authors: Xiao Xian Li, Wei Jie Li, Liu Yun
• Keywords: Heat Assessment, HowNet, Micro-Blog, Topic Detection

References

[1] Liu, Qing-Lei: Researches of Chinese sentence similarity based on HowNet (International Conference on Apperceiving Computing and Intelligence Analysis 2010).

DOI: 10.1109/icacia.2010.5709843

[2] Li, Wen: Fine-grained sentiment classification based on HowNet (Journal of Convergence Information Technology 2012).

[3] Huang, Sheng: News topic detection based on hierarchical clustering and named entity (NLP-KE 2011 - Proceedings of the 7th International Conference on Natural Language Processing and Knowledge Engineering 2011).

DOI: 10.1109/nlpke.2011.6138209

[4] Pullar, R.C.: Method for the preparation of aligned fibre samples for magnetic measurement using VSM (Journal of Magnetism and Magnetic Materials 2000).

DOI: 10.1016/s0304-8853(00)00350-4

[5] Zhang, Dan: Topic detection based on K-means (2011 International Conference on Electronics, Communications and Control).