amazon-product-reviews-sentiment-analysis
Sentiment analysis of product reviews in positive and negative with Identification of most reviewed products from Amazon products dataset.
-
Out of 37000 reviews, getting most reviewed products with number of reviews for each.
-
Do a sentiment analysis on how many reviews are positive and negative based on ratings given by the user.
-
Sentiment of reviews were not given beforehand in the dataset, so I took ratings in consideration and marked them Positive (>=3) and Negative (<3).
Installation
-
Install Jupyter Notebook and Python 3.7
-
Install Python libraries mentioned in requirements.txt for this project
-
On top of all the requirements mentioned in requirements.txt, follow installation of Huggingface Transformers on your system as sentiment analysis is being done using the transformer.
-
Click here for installation instructions:
https://huggingface.co/transformers/installation.html
Working with Jupyter Notebook
- Open notebook (.ipynb) from this project with Jupyter Notebook