Electronic Opinion Analysis System for Library (E-OASL)

Abstract

This paper demonstrates a new algorithm for Electronic Opinion Analysis System (E-OASL) for university library where it can analyze the user opinion on library and categorized it in different classification such as positive, negative, neutral and suggestions. The system also shows the percentage of each classification as system output. Our proposed algorithm is based on hybrid approach of sentiment analysis where machine learning's rule-based classifier and lexicon approach's corpus datastore are utilized. We needed to collect around 1200 raw data from the different types of user who are using university library to build our algorithm. The system integrates MySQL database for faster and better data processing. The paper also shows the evaluation of the E-OASL which showed satisfactory result of effectiveness and efficiency of EOASL compare to manual opinion analysis approach.