Flask NLP Ensemble Classifier

This repository contains a small Flask app that loads pre-trained models and TF-IDF vectorizers from a model/ directory and serves a web UI for text classification.

Expected model files (place into model/):

• logistic regression model (e.g. logistic_model.pkl or similar)
• LightGBM model (e.g. lgbm_model.pkl)
• CatBoost model (e.g. catboost_model.pkl)
• TF-IDF vectorizers: word-level (e.g. tfidf_word.pkl) and char-level (e.g. tfidf_char.pkl). If only one is present, it will be used for both.

Files added:

• app.py — Flask application entrypoint
• utils.py — helpers: load models, clean text, vectorize and ensemble predictions
• templates/index.html — frontend template
• static/style.css — basic styling
• requirements.txt — Python dependencies

How to run:

1. Create a virtual environment and install requirements:

   python -m venv myenv; .\myenv\Scripts\activate; pip install -r requirements.txt

2. Place your trained models and vectorizers into the model/ folder (see expected names above).

3. Run the app:

   python app.py

4. Open http://127.0.0.1:5000 in your browser.

Notes:

• The app cleans input text, vectorizes with TF-IDF (word and char), obtains probabilities from the three models, averages them, and shows Positive (>= 0.5) or Negative (< 0.5).
• If any model fails to load or predict, the app will skip it. Ensure models implement predict_proba for proper probabilities.