Automatic sentiment classification of political news articles.

• Main Author: Nourbakhsh, Armineh.
• Other Authors: Jin Cheon Na
• Format: Theses and Dissertations
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Document and text processing
• Online Access: http://hdl.handle.net/10356/18718
• Institution: Nanyang Technological University

The purpose of this study is to develop an automatic method for sentiment analysis of political news articles. The study analyses expressions of sentiment, opinion and emotion in political news articles in order to develop a lexicon of biased words and phrases. An automatic classifier is formulated that can identify biased words and phrases with the help of the dictionary, and predict the sentiment of various appraisers toward various objects in sentences. Sentiment is a person’s feeling, emotion or attitude toward an object. It is a multi-faceted aspect of textual expressions. Sentiment and appraisal might refer to the emotion or attitude of the author of the text, the polarity of the text (i.e. whether it is positive or negative), or the way sentiment is expressed (e.g. by emotional or critical statements). Sentiment is one of the non-topical aspects of text that is of interest to psychology, sociology, business, and management researchers. Business organisations analyse their customers’ sentiment toward their products for feedback analysis and business intelligence purposes. Communication researchers analyse sentiment in spoken and written media in order to investigate its impact on communication. One major source of data for communication researchers is political news articles. Researchers investigate public opinion about various issues by analysing sentiment in political news. They also analyse how media represents various politicians in order to investigate its impact on the approval the politician receives from the public. Manual sentiment analysis requires tremendous amount of time, cost, and human effort. Studies on automatic sentiment analysis attempt to reduce this cost by employing statistical or machine-learning models for analysing sentiment in text. Most models make use of a two-stage process: selecting the features of text that indicate sentiment and constructing a statistical or machine learning model that uses these features to predict the type of sentiment in text.