Sentimental Analysis 📊

A sentimental analysis is on reviews of phones sales of Amazon. Over 400k+ rows of reviews are proccesed to achieve an accuracy of 89.7% using logistic regression.

Dataset

Dataset used : Amazon Phone Sales Dataset

• 400k+ columns of reviews

Processure

• Starting from data cleaning, filtering out NaN columns for consistency.
• TF-IDF (Term Frequency-Inverse Document Frequency) used for feature extraction.
• Logistic Regression in order to classify between positive or negative reviews.

Working

Write a simple review focused on a phone and submit to get results.

Demonstration

Note

In order to alter the model, check the .ipynb file and save new model locally to run new results.