Exploiting Artificial Immune System to Optimize Association Rules for Word Sense Disambiguation

Abstract

Requirement specification is the major activity in software development. Since requirements are gathered from the customers in natural languages they are prone to ambiguities. Ambiguous requirements give many interpretations of the same word or sentence. In order to reduce the problems faced due to requirement ambiguities, many techniques have been proposed in the past. Word Sense Disambiguation is a bottleneck in most of the Natural Language Processing (NLP) applications. The approaches to deal with WSD aims to provide the best possible meaning for the target word which is lexically ambiguous. To reduce the ambiguities and optimize the association mining rules for Word Sense Disambiguation (WSD), this paper proposes a new approach based on Artificial Immune System and Association Rule Mining. The approach shows significant results when tested on a collection of many Software Requirement Specifications (SRS) Documents. The average accuracy provided by the system is 89.2725%. outperforms state of the art methods.