This project is a simple implementation of sentiment analysis using the Support Vector Machine (SVM) algorithm with scikit-learn. The dataset is sourced from a public GitHub repository, and the model utilizes TF-IDF vectorization to convert text into numerical features.

main.ipynb: The main notebook containing the entire process of training, evaluation, and prediction of the model.

vectorizer.sav: The saved vectorizer model file after training.

classifier.sav: The trained SVM classification model file.

README.md: Documentation of this project.

Python 3.x

Pandas

Scikit-learn

TfidfVectorizer

Pickle

Jupyter Notebook

Clone this repository:

Open the notebook:

Run each cell one by one to see the results.

Preprocessing: Convert text to numerical features using TF-IDF.

Training: Using SVM with a linear kernel.

Evaluation: Generates classification report including accuracy, precision, and recall metrics.

Prediction: Predict sentiment labels from manual text input.

Label Precision Recall F1-score Support Positive 0.919 0.910 0.915 100 Negative 0.911 0.920 0.915 100

This model demonstrates balanced performance in classifying positive and negative reviews, with an average F1-score of around 0.915.

162: pos - "odin is a great high school basketball ..." 600: pos - "bob the happy bastard's quickie review..." 659: pos - "good films are hard to find these days..." 1099: neg - "so ask yourself what '8mm' is all about..." 337: pos - "will hunting (matt damon) is a genius..."

Review: SUPERB, I AM IN LOVE IN THIS PHONE Prediction: pos

Review: Do not purchase this product. My cell phone blast when I switched the charger Prediction: neg

Review: I received defective piece display is not working properly Prediction: neg

Review: It's not even 5 days since I purchased this product. I would say this a specially blended worst Phone in all formats. Prediction: neg

This project is created solely for educational and exploratory purposes. Feel free to use it with proper attribution.