Konkani SentiWordNet: Resource for sentiment analysis using supervised learning approach

Abstract

Sentiment Analysis (SA) is the process of analyzing and predicting the hidden attitude/opinion in the given text expressed by an individual. Till now, ample amount of work has been carried out for the English language. But, no work is performed for the language Konkani in the field of Sentiment Analysis. Lexicon-based SA is a good beginning for any language, especially if the digital content is limited. Hence, the main motive of this paper is; to present the sentiment lexicon called SentiWordNet for Konkani language. The process of creating Konkani SentiWordNet is under progress using the Supervised Learning Approach. In this approach, the training set is generated using a Synset Projection Approach and Support Vector Machine (SVM) algorithm to classify the data. The reason behind using the Synset Projection Approach for building a training dataset is; English Sentiwordnet is developed using Semi-Supervised Approach where the training dataset is generated using WordNet lexical relations but; in Konkani WordNet, lexical relations are not yet developed. Hence, Synset Projection Approach is preferred. Conducted experimental results for the proposed algorithm are reported in this paper.