A Dynamic Weight Assignment Approach for Index Terms

Abstract

Currently in the development of Informational Retrieval (IR) systems, a set of keywords represents the semantics of a text document and after assigning weights to the set of keywords, they are used for indexing, searching and retrieval purposes. The current approaches for assigning weights are provided by an IR model that is used by an IR system. The main objective of the weight assignment was to provide ranking feature to an IR system. The retrieval performance of an IR system mainly depends on two parameters: extraction of a good set of keywords from text documents and the use of a good weight assignment approach. Most of currently available weight assignment approaches do not suggest any change to the weights of keywords after their initial assignment. It means that these approaches are static. In this study, we propose a dynamic weight assignment approach for weight assignment. In our opinion, using this proposed approach can be helpful in improving the retrieval performance of an IR system. This approach can be used as part of any IR model after initial assignment of weights.