NLP-With-Amazon-Food-Reviews

Problem Statement

While shopping online, customers widely look for reviews and ratings(1-5) for the products, which could help in understanding the reliability of the product before they initiate a purchase.

Key questions to address after completion of this project:

• Is it possible to determine the sentiment of the review?
• What words tend to indicate positive and negative reviews?
• What kind of reviews tend to be more helpful?
• How is the word count related to sentiment of a review?

Assumption

• Given a product review, using machine learning models to determine whether the sentiment of the review is positive or negative.
• Use F1 Score, which balances precision and recall, as validation metrics.

Data

• Amazon Fine Food Reviews https://www.kaggle.com/snap/amazon-fine-food-reviews
• This dataset consists of reviews of fine foods from amazon. The data span a period of more than 10 years, including all ~500,000 reviews up to October 2012.
• Reviews include product and user information, ratings, and a plain text review. It also includes reviews from all other Amazon categories.

Modelling

• Ensemble models:

  • Random Forest
  • Gradient Boosting
  • XGBoost

• Neural Network Models:

  • Convolutional Neural Network
  • Recurrent Neural Network

Structure of the code

1. Data Exploratory & Visualization
2. Data Cleaning and Feature Engineering (Encoding)
3. Random Forest & Gradient Boosting & RNN
4. XGBoost & CNN

Reference

1. TEXT PREPROCESSING USING PYTHON
   https://www.kaggle.com/shashanksai/text-preprocessing-using-python
2. Practical Text Classification With Python and Keras
   https://realpython.com/python-keras-text-classification/#your-first-keras-model
3. Exploratory Data Analysis
   https://www.kaggle.com/kanavbansal/exploratory-data-analysis