The document discusses various techniques for sentiment analysis, including feature selection techniques like pointwise mutual information (PMI), chi-square, TF-IDF, and association rule mining. It provides examples of features that can be extracted from text like names, terms, and relations. PMI is discussed in more detail, including how it is used to find collocations in text based on the information gained from the co-occurrence of words. Code in Python is also provided as an example of how to calculate PMI for bigrams and trigrams.