Homophone Ambiguity Reduction from Word Level Speech Recognition using Artificial Immune System

Nowadays Speech Recognition is a commanding intermediate that assists in Man-Machine Interfacing. Most of the conventional Speech Recognition Techniques use lexical analysis approach to comprehend a correct homophone, but still, it is a challenging area in word-level speech recognition. The current research proposes unsupervised reinforcement learning based Artificial Immune Algorithm to decrease the homophone ambiguity from any existing speech to text techniques in word-level speech recognition. It also applies an Affinity Maturation technique to premature the homophone dataset, collected from different sources and merged into one dataset, earlier before applying the proposed Artificial Immune System Algorithm for the better result at preliminary phase.