SENTIMENT POLARITY WITH SENTIWORDNET AND MACHINE LEARNING CLASSIFIERS

Akshaya.R Garje, Karbhari V. Kale

Abstract


: Sentiment classification is concerned with using automated methods for predicting the orientation of subjective content on textual content documents, with applications on some of areas consisting of recommended and advertising and marketing systems, customer intelligence and information retrieval. SentiWordNet is not anything however an opinion lexicon derived from the WordNet database in which each term is related to numerical scores indicating their sentiments. This research offers the results of making use of the SentiWordNet lexical resource to the hassle of computerized sentiment classification on labelled dataset. Our method incorporates counting positive and negative scores to decide sentiment orientation, and an improvement is provided by means of constructing a information set of applicable features using SentiWordNet as supply, and additionally implemented to a machine learning classifier. We compared the accuracies results obtained with SentiWordNet and Machine Learning Classifiers.

Keywords


Sentiment Analysis, SentiWordNet, Naïve Bayes, Support Vector Machines.

Full Text:

PDF


DOI: https://doi.org/10.26483/ijarcs.v8i9.5174

Refbacks

  • There are currently no refbacks.




Copyright (c) 2017 International Journal of Advanced Research in Computer Science