Skip to content

README.md

Latest commit

 

History

History
42 lines (35 loc) · 1.65 KB

README.md

File metadata and controls

42 lines (35 loc) · 1.65 KB

📢 Amazon Alexa Reviews Sentiment Analysis

This project performs Natural Language Processing (NLP) on Amazon Alexa reviews to classify sentiments using Decision Tree and Random Forest models. It includes data preprocessing, visualization, and model training with Flask integration for deployment.

🚀 Features

✔️ Sentiment analysis of Amazon Alexa reviews
✔️ Data visualization using Matplotlib & Seaborn
✔️ Word cloud representation of positive & negative words
✔️ Preprocessing with stop words removal & stemming
✔️ Model training using Decision Tree & Random Forest
✔️ Hyperparameter tuning with GridSearchCV
✔️ Flask-based web interface for predictions

📂 Dataset

The dataset contains Amazon Alexa reviews with:

  • verified_reviews (Text of the review)
  • rating (Star rating of the review)
  • feedback (1 for positive, 0 for negative)
  • variation (Alexa device variant)

📊 Exploratory Data Analysis

  • Distribution of ratings & feedback
  • Variation-wise rating analysis
  • Word clouds for positive & negative words

🛠️ Tech Stack

  • Python
  • Flask
  • NLTK (Text preprocessing)
  • Scikit-learn (Machine Learning)
  • Matplotlib & Seaborn (Data Visualization)
  • Pandas & NumPy (Data Manipulation)
  • Pickle (Model saving)

🔧 Model Training

  1. Preprocessing: Tokenization, Stopword removal, Stemming
  2. Feature Extraction: CountVectorizer (Bag of Words)
  3. Training:
    • Decision Tree Classifier
    • Random Forest Classifier
  4. Hyperparameter Tuning: GridSearchCV