🧠 AI Requirement Classifier

An intelligent NLP-powered web application that automatically classifies software requirements into structured categories — saving time, reducing manual effort, and improving requirement engineering workflows.

---

🚀 Live Demo

👉 Launch App

---

📌 Overview

Software requirement classification is a critical but tedious step in requirement engineering. This project uses Natural Language Processing (NLP) and a Naive Bayes classifier trained on labeled requirement data to instantly predict the type of any given requirement.

What it classifies:

Code	Category	Description
FR	Functional Requirement	What the system must do
NFR	Non-Functional Requirement	How the system must perform
PE	Performance Requirement	Speed, throughput, response time
US	Usability Requirement	Ease of use and accessibility
A	Assumption	Conditions assumed to be true
L	Limitation	Explicit system constraints
PO	Policy Requirement	Legal, regulatory, or organizational rules

---

✨ Features

• ⚡ Real-time prediction — instant classification as you type
• 📊 Confidence score — shows model certainty for each prediction
• 📘 Full explanations — each category includes a plain-English description
• 🎨 Clean interactive UI — built with Streamlit, no setup required for end users
• 🔤 NLP-based — uses TF-IDF vectorization for robust text understanding
• 📂 Category reference panel — all 7 categories always visible on screen

---

🛠️ Tech Stack

Component	Technology
Language	Python 3.8+
ML Model	Naive Bayes (MultinomialNB)
Vectorizer	TF-IDF (TfidfVectorizer)
Label Encoding	LabelEncoder (scikit-learn)
Web Framework	Streamlit
Model Persistence	Pickle

---

⚙️ How It Works

User Input (text)
      │
      ▼
TF-IDF Vectorizer  ──→  Converts text into numerical feature vector
      │
      ▼
Naive Bayes Model  ──→  Predicts requirement category
      │
      ▼
Label Encoder      ──→  Maps prediction back to human-readable label
      │
      ▼
Result displayed with category name, code, description & confidence score

1. User types or pastes a software requirement
2. Text is preprocessed and converted into a TF-IDF feature vector
3. The trained Naive Bayes model predicts the category
4. The label encoder maps the numeric prediction back to a category code
5. The UI displays the full category name, description, and confidence score

---

📂 Project Structure

AI-Requirement-Classifier/
│
├── app.py                  # Main Streamlit application
├── re_nb_model.pkl         # Trained Naive Bayes model
├── tfidf_vectorizer.pkl    # Fitted TF-IDF vectorizer
├── label_encoder.pkl       # Fitted label encoder
├── requirements.txt        # Python dependencies
└── README.md               # Project documentation

---

▶️ Getting Started

Prerequisites

• Python 3.8 or higher
• pip

Installation & Run

# 1. Clone the repository
git clone https://github.com/your-username/AI-Requirement-Classifier.git

# 2. Navigate into the project folder
cd AI-Requirement-Classifier

# 3. Install dependencies
pip install -r requirements.txt

# 4. Launch the app
streamlit run app.py

The app will open automatically at http://localhost:8501

Requirements.txt (minimum)

streamlit
scikit-learn
numpy

---

📊 Example Inputs & Outputs

Requirement Text	Predicted Category
"The system shall allow users to log in using email and password."	✅ FR — Functional Requirement
"The system must respond to all requests within 2 seconds under normal load."	✅ PE — Performance Requirement
"The UI shall be operable by users with no technical background."	✅ US — Usability Requirement
"The system shall be available 99.9% of the time."	✅ NFR — Non-Functional Requirement
"We assume all users will have internet access."	✅ A — Assumption
"All data storage must comply with GDPR regulations."	✅ PO — Policy Requirement
"The system does not support offline mode."	✅ L — Limitation

---

🧪 Model Details

• Algorithm: Multinomial Naive Bayes — well-suited for text classification tasks
• Vectorization: TF-IDF (Term Frequency–Inverse Document Frequency) captures word importance relative to the dataset
• Training data: Labeled software requirement dataset with 7 categories
• Persistence: Model, vectorizer, and encoder saved as .pkl files using Python's pickle module for fast loading at runtime

---

🗺️ Roadmap

• Real-time single requirement classification
• Confidence score display
• Category reference panel
• Bulk classification via CSV upload
• Model accuracy improvements (SVM, BERT)
• Export results to CSV / PDF
• REST API endpoint for integration
• Support for more requirement categories
• Dark / light mode toggle

---

👩‍💻 Developed By

Huma
AI & Machine Learning Enthusiast 🚀
Built as part of an NLP & Machine Learning project

---

🤝 Contributing

Contributions, issues, and feature requests are welcome!

1. Fork the repository
2. Create a feature branch (git checkout -b feature/your-feature)
3. Commit your changes (git commit -m 'Add your feature')
4. Push to the branch (git push origin feature/your-feature)
5. Open a Pull Request

---

📄 License

This project is licensed under the MIT License — feel free to use, modify, and distribute.

---

⭐ If you found this project helpful, please give it a star on GitHub — it means a lot!