Sentiment Analysis of ISIS Related Tweets Using Absolute Location

Twitter is a free broadcast service for the registered members to the public limited to 140 characters that may include text, photos, videos and hyperlinks. People share news, opinions and information to support or against media. The most petrified topic is the ISIS terrorist attacks taking place around the world. ISIS takes advantage of the social media to continuously communicate using coded words or to establish their indirect presence. Hashtags associated with ISIS can be analyzed and capture the sentiment of the tweets. This paper presents a novel process for sentiment analysis on the ISIS related tweets and to organize the opinions with their geolocations. The Jeffrey Breen algorithm is used for sentiment analysis. The data mining algorithms such as Support Vector Machine, Random Forest, Bagging, Decision Trees and Maximum Entropy are applied for polarity based classification of ISIS related Tweets. The results are compared and presented.