News Articles Category Prediction

[pratikwayal01]

Pull Request Title:

News Articles Category Prediction

Issue Number #108

Description:

This pull request implements the following changes:

• Added a Support Vector Classifier (SVC) model to the Streamlit app for classifying news articles into categories.
• Integrated GridSearchCV to optimize hyperparameters for the SVC model, improving classification accuracy.
• Updated the Streamlit interface to allow users to input text and classify it using the best-performing model (SVC).
• Modified requirements.txt to include necessary dependencies such as scikit-learn, streamlit, and other relevant libraries.
• Enhanced the README.md file with a detailed project description, setup instructions, and usage notes.

Changes Made:

• model.py: Created a new file for building and exporting the SVC model with the optimized hyperparameters.
• app.py: Integrated the SVC model into the Streamlit app, allowing the user to classify news articles.
• requirements.txt: Added required libraries such as scikit-learn, joblib, and others.
• README.md: Updated the README file with more comprehensive instructions, including the description of the Jupyter Notebook used for model building.

Testing:

• Tested the Streamlit app locally to ensure that the SVC model correctly classifies articles based on input.
• Verified that the app responds as expected with various inputs and handles edge cases.
• Ran cross-validation on the SVC model, confirming that the accuracy has improved compared to the baseline model.

[yashasvini121]

Please follow these instructions:

• Store the notebook in the notebooks/ directory.
• Update the existing requirements.txt instead of creating a new one.
• Save the pickle file in the saved_models/ directory.
• Ensure the model_details function is implemented.
• Replace app.py with a page using page_handler. Refer to the docs/ for detailed steps.
• Delete the readme.md file.
• [[IMPORTANT]] Name the function that is used for prediction as get_prediction

For better clarity, refer docs/.
Hope this helps.