Document how CountVectorizer works

[farukcankaya]

We are using CountVectorizer from from sklearn.feature_extraction.text in Milestone_1_W_Relevant_Data and Milestone_1 like below:

count_vectorizer = CountVectorizer(analyzer='word',       
                             min_df=3,                       
                             stop_words='english',             
                             lowercase=True,                   
                             token_pattern='[a-zA-Z0-9]{3,}',  
                             max_features=5000,          
                            )

• Dig into this vectorizer to understand well how it is working. Specifically understand, and document, and prepare code examples for each parameter of the vectorizer.
• How important to the selected number of max_features?
• Could we give a stopword list manually to the vectorizer? i.e. union of accounting specific stop word list and standard English stop wordlist