Sentiment classification

In sentiment classification, one message can be classified as either positive or negative. This is excellent to get insight about how the public think about a person, products, or services.

In this chapter, we will classify the tweets to get a personal positive or negative feeling. It is important to clarify that tweets are limited to 140-characters length with a very casual language and in many cases the message may be very noisy with usernames, links, repeated letters, and emoticons. However, Twitter provides a way to get feedback about large amount of topics in real-time. We can see sample tweets as follows:

"Photoshop, I hate it when you crash " - Negative
"@Ms_HipHop im glad ur doing weeeell " - Positive

The general process of the sentiment classification is presented in the following screenshot. We start extracting the features (words) from the training data (Text Corpus). Then, we need to train the classifier with a bag of words, which is a list of words and...