Sentiment Analysis

This project focuses on sentiment analysis using a Twitter dataset consisting of 50,000 tweets. The dataset is split into training and testing sets, with 80% used for training and 20% for testing. The sentiment analysis model achieved an accuracy of 90%. The alogorithm which I have used is RandomForest along with vectorization and preprocessing containing in a single machine learning pipeline.

Note: If the deployed link is not working due to error in resources kindly prefer running locally using command mentioned below.

Project Structure

• data_process/: It is custom made python module which consists of all the neccessary function for data preprocessing pipeline.
• main.ipynb: Jupyter notebook file with training history and results.
• requirements.txt: File listing the required dependencies for the project.
• app.py: File which contains the neccessary coding for the frontend interaction using streamlit.
• sentiment_analysis.pkl: It is exported trained model after training

Getting Started

1. Clone the repository:

   git clone https://github.com/dhruvk2002/sentiment-analysis.git
   

2. Navigate to the Project Directory

   cd sentiment-analysis
   

3. Install Dependencies

   pip install -r requirements.txt
   

4. Run the streamlit app

   streamlit run app.py