🧠 AI – Intent Based Chatbot (PyTorch + Tkinter)

A simple desktop chatbot built using Natural Language Processing (NLP) and PyTorch.
This chatbot understands user input, predicts the intent using a neural network, and responds accordingly through a clean Tkinter GUI.

---

📌 Project Overview

This project is an intent classification chatbot.

It works by:

1. Converting user text into numerical format (Bag of Words)
2. Passing it through a Neural Network
3. Predicting the most likely intent
4. Returning a predefined response

The goal of this project was to understand:

• How chatbots process text
• How machine learning models classify text
• How neural networks work in real applications
• How to connect ML models with a GUI

---

🛠 Technologies Used

• Python 3
• PyTorch (Neural Network)
• NLTK (Tokenization & Stemming)
• Tkinter (GUI)
• NumPy

---

🧠 How It Works (Simple Explanation)

1️⃣ Text Preprocessing

• Tokenization (splitting sentence into words)
• Lowercasing
• Stemming (reducing words to root form)
• Removing punctuation

---

2️⃣ Bag of Words

Each sentence is converted into a numerical vector.

Example:

Input:

"Hello how are you?"

Converted into:

[0, 1, 0, 1, 0, ...]

This allows the neural network to understand text mathematically.

---

3️⃣ Neural Network Architecture

• Input Layer → Bag of Words size
• Hidden Layer 1 → ReLU
• Hidden Layer 2 → ReLU
• Output Layer → Number of intents

Loss Function: CrossEntropyLoss
Optimizer: Adam

The output layer gives probability scores for each intent.

---

4️⃣ Intent Prediction

The model:

• Predicts the highest probability intent
• If confidence > 75%, it returns a response
• Otherwise, it says it doesn’t understand

---

🖥 GUI Interface

The chatbot runs inside a simple Tkinter desktop window.

Features:

• Clean student-level UI
• User & Bot message colors
• Send button + Enter key support
• Scrollable chat window

---

🚀 How To Run

1️⃣ Install dependencies

pip install torch nltk numpy

2️⃣ Train the model

python train.py

3️⃣ Run the GUI

python gui.py

---

📂 Project Structure

SMART-AI-Chatbot/
│
├── train.py
├── gui.py
├── model.py
├── nltk_utils.py
├── intents.json
├── data.pth
└── README.md

---

WORKING Screenshot

---

🎯 What I Learned From This Project

• How text is converted into numbers
• How neural networks classify text
• Understanding of input size, hidden layers, and output layers
• Using PyTorch for training models
• Connecting ML models with a GUI
• Debugging real ML errors

---

📌 Future Improvements

• Add more intents
• Use advanced NLP models (like Transformers)
• Add memory/context support
• Improve GUI design
• Deploy as web application

---

📖 Learning Note

This project was built as a learning exercise by referencing tutorials and understanding each component step-by-step.
The goal was not just to copy, but to understand how chatbot systems work internally.

---

👨‍💻 Author

Built as part of my Machine Learning & NLP learning journey.