Selecting effective index terms using a decision tree

This paper explores the effectiveness of index terms more complex than single words used in conventional information retrieval systems. Retrieval is performed in two phases. In the first phase, a conventional retrieval method (the Okapi system) is used and in the second phase, complex index terms such as syntactic relations and single words with part of speech information are introduced to rerank the results of the first phase. The effectiveness of the different types of index terms were evaluated through experiments, in which the TREC-7 test collection and 50 queries were used. The experiments showed that retrieval effectiveness was improved for 32 out of 50 queries. Based on this investigation, we introduced a method to select effective index terms by using a decision tree. Experiments with the same test collection showed that retrieval effectiveness was improved in half of