prashanth-nani / sentiment-analysis Goto Github PK

It is a system built using Python and NLTK to analyse the reviews of a product obtained from e-commerce websites. The reviews are first cleaned using suitable regular expressions, stemming, removal of stop words and word tokenizing. Thereafter they are classified using Naive Bayes to determine the sentiment as being positive, negative or neutral.

The reviews are scraped from snapdeal website. The features of this tool are as follows.

• Option to choose a proxy connection
• Works with authenticated proxy
• Fetches all the reviews of ALL the mobile phones available
• Record logs to capture errors
• Manages directories by itself (Automatic cleaning up of files when execution starts)

• Python 3.4+
• Python modules:
  • bs4
  • requests (v2.9.1+)
  • os
  • re

$ sudo apt-get install python3
$ sudo apt-get install python3-pip  # Need atleast v8.0.3

If pip3 --version is <8.0.3. You can get it here and install it by running sudo python3 get-pip.py. Then continue with the remaining installation process.

$ sudo pip3 install bs4
$ sudo pip3 install requests
$ sudo pip3 install regex

python3 snapdeal.py #Uses configured python 3+ version available