In this section, we will learn how to analyze documents quantitatively. A simple way is to look at the distribution of unigram words across the document and their frequency of occurrence, also termed as term frequency (tf). The words with higher frequency of occurrence generally tend to dominate the document.

However, one would disagree in case of generally occurring words such as the, is, of, and so on. Hence, these are removed by stop word dictionaries. Apart from these stop words, there might be some specific words that are more frequent with less relevance. Such kinds of words are penalized using their inverse document frequency (idf) values. Here, the words with higher frequency of occurrence are penalized.

In this section, we will generate a tf-idf matrix across chapters of the book Pride and Prejudice.