Tracking dynamics of topic trends using a finite mixture model
Proceedings of the tenth ACM SIGKDD international conference on Knowledge discovery and data mining
The predictive power of online chatter
Proceedings of the eleventh ACM SIGKDD international conference on Knowledge discovery in data mining
Generative model-based document clustering: a comparative study
Knowledge and Information Systems
Incorporating with Recursive Model Training in Time Series Clustering
CIT '05 Proceedings of the The Fifth International Conference on Computer and Information Technology
Short communication: Variable space hidden Markov model for topic detection and analysis
Knowledge-Based Systems
Hi-index | 0.00 |
Hierarchical topic structure can express topics in a natural way which is more reasonable for human machine interface. However, the hierarchical topic structure that is extracted by most of the topic analysis algorithms can not present a meaningful description for all subtopics in the hierarchical tree. We propose a new hierarchical clustering algorithm based on variable feature selection for each level in the hierarchical structure. The algorithm employs a top-down strategy to extract subtopics and setups the relation between topics in neighbor levels based on common documents number. The number of the levels in the hierarchical structure is determined by the frequency of the selected word feature. Experiments on a real world dataset which is collected from a news website shows that the proposed algorithm can generate more meaningful topic structure, by comparing to the current hierarchical topic clustering algorithms.