Sentiment Analysis Based On Context Dependent Words

One of the important types of information on the Web is the opinions expressed in the user generated content, e.g., customer reviews of products, forum posts, and blogs etc. In this thesis paper, user reviews of products have been focused on. In particular, the problem of determining the semantic orientations (positive, negative or neutral) of opinions expressed on product features in reviews is studied. This problem has many applications, e.g., opinion mining, summarization and search.Following work has been finished in this research:(1) Here in this thesis, traditional learning methods for sentiment analysis are examined and compared them against each other, such as naive Bayesian, support vector machines(SVM), and K-means.(2) Our approach has been presented as well, to compute semantic orientation of a sentiment bearing word. Opinion words are words that express desirable (e.g., great, amazing, etc.) or undesirable (e.g., bad, poor, etc) states. Heuristics has been applied on certain predefined predicates expressing semantic relationship between two concepts for classifying words that have a positive or negative semantic orientation and finding words that have similar semantic orientation.(3) Study on the treatment of opinion word extraction, using some semantic rules and mutual information method and other pretreatment technology to do word feature extraction.(4) Our approach was tested on publicly available dataset and the results of our experiments are presented and tradeoffs and limitations of the proposed solution are also examined. It is possible to determine semantic orientation of words with high accuracy by exploiting a machine-understandable layman’s knowledge and basic facts that ordinary people know about the world, is the basis of paper’s conclusion.