Measuring semantic similarity in WordNet by using Neural Network and Differential Evolution Algorithm

Semantic similarity between two words is an important problem in many applications of information retrieval and word sense disambiguation. In this paper, we calculate semantic similarity between two words by using the lexical database called WordNet and neural network. The neural network learning process is formulated as an optimization problem and optimized by using the Differential Evolution algorithm. We use the Rubenstein and Goodenough, and Miller and Charles datasets to test the similarity results. Our model produces high values of correlations for both the datasets.