🌟 Chatbot with RASA and NLU Model 🌟

An AI-powered conversational assistant designed to revolutionize user interaction for E-commerce.

---

🎯 Business Objective

A chatbot serves as a digital assistant, engaging users in natural conversations—whether to fetch product details, provide support, or process requests seamlessly. This project focuses on building an AI-based chatbot using the Rasa NLU framework, enabling dynamic and intelligent communication.

Use Case: An E-commerce assistant capable of:

• 📦 Providing product information (product_info)
• 💰 Answering price inquiries (ask_price)
• ❌ Handling order cancellations (cancel_order)

---

🚀 Project Overview

What’s Inside?

• Intents: Understanding user goals (e.g., ask_price, cancel_order)
• Entities: Extracting contextual data (e.g., product, order_id, location)
• Pipeline: Built using Spacy and TensorFlow for robust natural language understanding.

Tech Stack 🛠️

• Language: Python 🐍
• Libraries: pandas, matplotlib, rasa, pymongo, tensorflow, spacy
• Database: MongoDB

---

📁 Folder Structure

🔹 Input:

Contains training data, configurations, and sample intents/entities:

• data.json
• spacy_config.yml
• tensorflow_config.yml

🔹 Src (Source Code):

The backbone of the project, containing modularized code:

• Engine.py: The main script orchestrating all functions.
• ML_Pipeline folder: Contains modular Python functions for:
  • Data preparation 📊
  • Model training 🤖
  • Evaluation metrics 📈

🔹 Output:

Pre-trained models for instant deployment. No need to retrain from scratch—just load and go! 🚀

🔹 Lib (Reference Materials):

Includes Jupyter notebooks, reference slides, and notes for deeper understanding.

---

🛠️ Key Features

1. Intent and Entity Recognition:
   • Captures user intent (product_info, ask_price) and extracts relevant entities (product, order_id).
2. Model Training Pipelines:
   • Supports Spacy and TensorFlow pipelines for intent classification and entity recognition.
3. Data Visualization & Insights:
   • Exploratory Data Analysis (EDA) for a deeper understanding of dataset patterns.
4. MongoDB Integration:
   • Efficiently manages session-based interactions.

---

📚 How It Works

1. 🧹 Data Preparation: Curate datasets from tools like Rasa NLU Trainer or Chatito.
2. 🧩 Modular Code: Functions are neatly organized for clarity and scalability.
3. 🎛️ Model Configuration: YAML files for Spacy and TensorFlow pipelines.
4. 🏋️ Training: Models trained on annotated datasets for intent and entity recognition.
5. 📊 Evaluation: Confusion matrix plots to compare models and select the best one.
6. 🤝 Chatbot Testing: Seamless real-time testing for robust performance validation.

---

💡 Project Takeaways

By the end of this project, you’ll learn:

• The fundamentals of AI-based chatbots.
• How to configure pipelines with Rasa NLU, Spacy, and TensorFlow.
• MongoDB integration for chatbot sessions.
• Building modularized, scalable Python codebases.

---

🎉 Want to Explore?

Try it Out!

1. Clone this repo:

   git clone https://github.com/Vidhi1290/Chatbot-with-RASA-NLU-Model-and-Python.git

2. Install dependencies:

   pip install -r requirements.txt

3. Train the model:

   python src/engine.py

4. Test the chatbot:

   python src/test_chatbot.py

---

📞 Connect with Me!

Vidhi Waghela

• ✉️ Email: vidhiwaghela99@gmail.com
• 📞 Contact: +91 9152257810
• 🌐 GitHub | Kaggle
• 💼 LinkedIn
• 📸 Instagram
• 🐦 X (Twitter)
• ✍️ Medium

---