1. Learn
    2. /
  2. Courses
    3. /
  3. Sentiment Analysis in Python

Exercise

BOW using product reviews

You practiced a BOW on a small dataset. Now you will apply it to a sample of Amazon product reviews. The data has been imported for you and is called reviews. It contains two columns. The first one is called score and it is 0 when the review is negative, and 1 when it is positive. The second column is called review and it contains the text of the review that a customer wrote. Feel free to explore the data in the IPython Shell.

Your task is to build a BOW vocabulary, using the review column.

Remember that we can call the .get_feature_names() method on the vectorizer to obtain a list of all the vocabulary elements.

Instructions

100 XP
  • Create a CountVectorizer object, specifying the maximum number of features.
  • Fit the vectorizer.
  • Transform the fitted vectorizer.
  • Create a DataFrame where you transform the sparse matrix to a dense array and make sure to correctly specify the names of columns.