https://github.com/chintan45/amazon-product-review-classification

A simple classifier to predict the sentiments of Amazon product reviews.
https://github.com/chintan45/amazon-product-review-classification

machine-learning nlp nltk sentiment-classification

Last synced: 12 days ago
JSON representation

A simple classifier to predict the sentiments of Amazon product reviews.

• Host: GitHub
• URL: https://github.com/chintan45/amazon-product-review-classification
• Owner: Chintan45
• Created: 2024-04-09T06:02:59.000Z (9 months ago)
• Default Branch: main
• Last Pushed: 2024-04-09T07:16:34.000Z (9 months ago)
• Last Synced: 2024-11-06T03:46:17.282Z (2 months ago)
• Topics: machine-learning, nlp, nltk, sentiment-classification
• Language: Jupyter Notebook
• Homepage:
• Size: 22 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Amazon Product Review Classification

- This project was focused on the development of a machine learning-based classifier designed to accurately categorize Amazon product reviews.

- By leveraging advanced algorithms such as Naive Bayes, SVM, Logistic Regression, XGBoost, and Random Forest, the classifier is able to achieve 89% accuracy and an 87% F1-score.

- In this project, I have used TF-IDF vectorizer for feature extraction to interpret and classify vast quantities of text data effectively.



### Project Overview

- **Objective**: To accurately classify Amazon product reviews using machine learning algorithms.

- **Algorithms Used**: Naive Bayes, SVM, Logistic Regression, XGBoost, and Random Forest.

- **Feature Extraction**: Utilized TF-IDF.

- **Validation Method**: 5-fold cross-validation.

- **Dataset**: 220,909 Amazon product reviews ([Link]('/AMAZON_FASHION_sample.json.gz')).

### Technologies Used

- **Python**: The primary programming language used for model development.

- **Jupyter Notebook**: For documenting the development process and performing data analysis.

- **Scikit-learn**: Utilized for model training, feature extraction, and evaluation.

- **XGBoost**: For implementing the XGBoost algorithm.

- **Pandas & NumPy**: For data manipulation and numerical calculations.

- **NLTK**: For text preprocessing and natural language processing tasks such as tokenization, stemming, lemmatization and parsing.

### Algorithm Performance

| Algorithm          | Accuracy (%) | F1-Score (%) |

|--------------------|--------------|--------------|

| Naive Bayes        | 88           | 83           |

| Random Forest      | 88           | 82           |

| XGBoost            | 88           | 85           |

| Logistic Regression| 89           | 86           |

| SVM                | 89           | 87           |