Comparison of Semantic Similarity Determination using Machine Learning

Evaluating the semantic similarity of two terms is a task central to automated understanding of natural languages. The challenge of “semantic similarity” lies in determining if two chunks of text have very similar meanings or totally different meanings. The amount of research on semantic similarity has increased greatly in the past 5 years, partially driven by the annual SemEval competitions. In this work, to compute the similarity between terms we consider the WordSim and SimLex data set, compare the results obtained between Neural network, Support Vector Machine and Linear Regression machine learning techniques and evaluate the results obtained against the M&C data set. For the data set considered, the Neural Network Model gave the best results, the Linear Regression method fared better than the Support Vector Machine with Regression