Skip to content

headless-start/cs2-endtoend-chatbot

Repository files navigation

Counter-Strike 2 Chatbot

📌 Project Overview

This project is an end-to-end Counter-Strike 2 Chatbot that provides real-time answers to user queries about the game. Built entirely using Flask, the chatbot allows users to ask questions about game economy, map strategies, skins, and more.

The backend leverages TF-IDF vectorization and Logistic Regression for intent classification, ensuring precise and relevant responses.


🚀 Key Features

  1. Flask-Powered Chat Interface:
    • The entire chatbot runs on Flask, handling both frontend and backend.
    • Users can type questions and receive instant responses.
    • Supports queries about game economy, map strategies, skins, and general gameplay tips.
  2. Intent Classification with Machine Learning:
    • Utilizes TF-IDF (Term Frequency-Inverse Document Frequency) for feature extraction.
    • Logistic Regression model classifies user queries into predefined categories.
    • Ensures fast and accurate response generation.

🔍 How It Works

  1. User Input:
    • Users enter their queries via the Flask-based chat interface (e.g., "What is the best economy strategy for Mirage?").
  2. Backend Processing:
    • The input text is processed using TF-IDF vectorization.
    • The Logistic Regression model predicts the intent of the query.
    • A relevant response is retrieved from the predefined knowledge base.
  3. Response Generation:
    • Flask sends the generated response back to the user via the chat interface.

🛠 System Requirements

Dependencies

  • Python 3.8+
  • Required Libraries:
    pip install flask nltk pandas scikit-learn
    

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.