Improving the Lexicon of Positive/Negative Words and Bigrams for Sentiment Analysis

Report published in the Proceedings of the National Conference on "Education in the Information Society", Plovdiv, May, 2013

Тhe idea of the current paper is to propose an algorithm for improving the list of positive and negative words based on a specific topic. The opinions are classified by a person or a machine (or combined) and the most frequent words are being found that are not in the list of stop words. These words are being added to the list and than with a sample classifier is found improvement in the classification of the already exctracted opinions. Several tests have been described to show how to test the algorithm.

Association for the Development of the Information Society, Institute of Mathematics and Informatics Bulgarian Academy of Sciences