Search engine For Twitter sentiment analysis

The purpose of sentiment analysis is to determine the attitude of a writer or a speaker with respect to some topic or his feeling in a document. Thanks to the rise of social media, nowadays there are numerous data generated by users. Mining and categorizing these data will not only bring profits for companies, but also benefit the nation. Sentiment analysis not only enables business decision makers to better understand customers' behaviors, but also allows customers to know how the public feel about a product before purchasing. On the other hand, the aggregation of emotions will effectively measure the public response toward an event or news. For example, the level of distress and sadness will increase significantly after terror attacks or natural disaster. In our project, we are going to build a search engine that allows users to check the sentiment of his query. Some of previous researches on classifying sentiment of messages on micro-blogging services like Twitter have tried to solve this problem but they have ignored neutral tweets, which will result in problematic results (12). Our sentiment analysis will also be based on tweets collected from twitter, since twitter can offer sufficient and real-time corpora for analysis. We will preprocess each tweet in the training set and label it as positive, negative or neutral. As we use words in the tweet as the feature for our model, different features will be used. We will show that accuracy achieved by different machine learning algorithms (Naïve Bayes, Maximum Entropy) can be improved with a feature vector obtained by using bigrams (5). In our practice, we find that Naive Bayes has better performance than Maximum Entropy.