Improving text categorization methods for event tracking
SIGIR '00 Proceedings of the 23rd annual international ACM SIGIR conference on Research and development in information retrieval
Capturing term dependencies using a language model based on sentence trees
Proceedings of the eleventh international conference on Information and knowledge management
Introduction to topic detection and tracking
Topic detection and tracking
Optimizing story link detection is not equivalent to optimizing new event detection
ACL '03 Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - Volume 1
A generative theory of relevance
A generative theory of relevance
An Improved Feature Selection using Maximized Signal to Noise Ratio Technique for TC
ITNG '06 Proceedings of the Third International Conference on Information Technology: New Generations
Relevance models for topic detection and tracking
HLT '02 Proceedings of the second international conference on Human Language Technology Research
Using Soundex codes for indexing names in ASR documents
SpeechIR '04 Proceedings of the Workshop on Interdisciplinary Approaches to Speech Indexing and Retrieval at HLT-NAACL 2004
| Hi-index | 0.00 |
Given two stories, Story Link Detection System identifies whether they are discussing the same event. Standard approach in link detection system is to use cosine similarity measure to find whether the two documents are linked. Many researchers applied query expansion technique successfully in link detection system, where models are built from the relevant documents retrieved from the collection using query expansion. In this approach, success depends on the quality of the information retrieval system. In the current research, we propose a new information retrieval system for query expansion that uses intra-cluster similarity of the retrieved documents in addition to the similarity with respect to the query document. Our technique enhances the quality of the retrieval system thus improving the performance of the Link Detection System. Combining this improved IR with our Cohesion Model provides excellent result in link detection. Experimental results confirm the effect of the improved retrieval system in query expansion technique.