Query Expansion and Evolution of Topic in Information Retrieval Systems

Approach based on clustering will be described in our paper. Basic version of our system was given in [5] allows us to expand query through special index. Hierarchical agglomerative clustering of the whole document collection generates the index. Retrieving of topic development is specific problem. Standard methods of IR does not allow us such kind of queries for appropriate solution of information problem. The goal of presented method is to find list of documents that are bearing on topic, represented by user-selected document, sorted with respect to historical development of the topic.

• Type: Conference
• Year: 2004
• Where: DATESO
• Authors: Jiri Dvorský, Jan Martinovic, Václav Snásel
• Keywords: Basic Version, DATESO 2004, DATESO 2007, Hierarchical Agglomerative Clustering, Topic Development