SENTIMENT ANALYSIS ON PUNJABI NEWS ARTICLES USING SUPPORT VECTOR MACHINE by Gagandeep Kaur
Abstract
Sentiment analysis is a field of Natural Language Processing and it is the most trending field of research. In the process of text mining that is used to find out people’s opinion about a particular product, topic and predicting market trends or outcomes of elections, detecting and classifying sentiments from the text. Sentiment analysis on the Punjabi language is to be performed because of increasing amount of Punjabi content over the web, provides an important aspect for the researchers, organizations, and governments to analyze the user-generated content and get the useful information from it. This work basically focuses on mining sentiments and analyzing them for the Punjabi language. With the increase in the amount of information being communicated via regional languages like Punjabi, comes a promising opportunity of mining this information. Nowadays, it is a new trend to read online news in a daily practice. People's opinion tends to be changed as per they read news content. The news content that they read normally about the negative content regarding various things for example rapes, corruption, thefts etc. Reading such negative news is spreading negativity around the society. So there is need to classify the positive and negative news content for creating a positive environment because if they read positive they think positive. Support Vector Machine approach is used by the proposed system to classify the content into different categories of news like crime, entertainment, politics, sports, and weather and then finding its polarity. The results of the proposed system depict remarkable accuracy. The accuracy of sentiment analysis on Punjabi news articles using Support vector machine is found to be 90%.
