Intent Classification Chatbot

Overview

This project implements an intent classification chatbot using various machine learning models. The system processes user input, classifies intents using ELMo embeddings, and selects the most appropriate response based on the predicted intent.

Project Structure

The project is structured as a single Python file (chatbot.py), which includes:

• Data preprocessing and analysis
• Model training and evaluation
• Chatbot implementation
• Flask API setup (commented out for future work)

Features

• Comprehensive EDA with visualizations
• Text preprocessing and cleaning
• ELMo embeddings for text representation
• Multiple machine learning models:
  • Support Vector Machine (SVM)
  • Random Forest
  • Gradient Boosting
  • Logistic Regression
• Model performance evaluation
• Interactive chatbot interface

Dependencies

All required packages are listed in requirements.txt. Install them using:

pip install -r requirements.txt

Data Requirements

The system requires an input file named dialogs.csv with the following columns:

• input: User messages
• output: Bot responses
• intent: Message intents

Usage

Data Analysis and Model Training

Run the following command to start data analysis and model training:

python chatbot.py

The script will:

• Perform EDA with visualizations
• Preprocess text data
• Train multiple models
• Save the best performing model
• Implement chatbot functionality

Model Training

The system trains and evaluates four different models:

• Support Vector Machine (SVM) (with grid search)
• Random Forest
• Gradient Boosting
• Logistic Regression

Each model is evaluated using the following metrics:

• Accuracy
• F1 Score
• Precision
• Recall
• Confusion Matrix

Output Files

• Trained model: best_logistic_regression_model.joblib
• Various visualization plots
• Performance metrics

Future Work

• Implementation of a Flask web interface
• Model deployment
• Response generation improvements
• Additional NLP features