Update README.md

README.md

@@ -1,19 +1,106 @@
 ---
-title: Anomoly Detection
-emoji: 🚀
-colorFrom: red
-colorTo: red
-sdk: docker
+title: NSL-KDD Anomaly Detection
+emoji: 🧠
+colorFrom: blue
+colorTo: green
+sdk: streamlit
 app_port: 8501
-tags:
-- streamlit
 pinned: false
-short_description: Streamlit template space
+tags:
+  - anomaly-detection
+  - streamlit
+  - machine-learning
+  - cybersecurity
+---
+
+# 🧠 NSL-KDD Anomaly Detection using Isolation Forest, One-Class SVM, and Logistic Regression
+
+This Streamlit web app performs **network intrusion / anomaly detection** using the **NSL-KDD** dataset — a popular benchmark for cybersecurity and anomaly detection research.  
+It compares three models:
+- 🏝 **Isolation Forest** (unsupervised)
+- 🔹 **One-Class SVM** (unsupervised)
+- ⚙️ **Logistic Regression** (supervised baseline)
+
+---
+
+## 🌐 Live Demo
+👉 Try it on **Hugging Face Spaces**:  
+*(Replace with your actual Space link after deployment)*
+
+---
+
+## 🧠 Overview
+
+The **NSL-KDD dataset** is an improved version of the classic **KDD Cup ’99** dataset for intrusion detection.  
+It contains labeled examples of **normal** and **attack** network connections, described by 41 features such as protocol type, service, flag, duration, and bytes transmitted.
+
+This project:
+- Loads the NSL-KDD dataset (from local file or online source)
+- Performs preprocessing and scaling
+- Trains and compares:
+  - Isolation Forest  
+  - One-Class SVM  
+  - Logistic Regression  
+- Computes accuracy, precision, recall, F1-score, and confusion matrices  
+- Visualizes anomalies using Seaborn and Matplotlib
+
+---
+
+## ⚙️ How It Works
+
+1. Load and clean NSL-KDD dataset  
+2. Convert categorical variables (e.g., protocol, service, flag) to numeric form  
+3. Scale the features using **StandardScaler**  
+4. Train and evaluate the following models:
+   - **Isolation Forest** → Detects outliers by random partitioning  
+   - **One-Class SVM** → Learns boundary around normal samples  
+   - **Logistic Regression** → Supervised baseline using true labels  
+5. Display results and metrics interactively in Streamlit
+
+---
+
+## 🧾 Dataset: NSL-KDD
+
+| Column Type | Description |
+|--------------|-------------|
+| Numerical | Continuous numeric features (e.g., duration, src_bytes, dst_bytes) |
+| Categorical | protocol_type, service, flag |
+| Target | normal / attack |
+
+📂 You can download it from:  
+https://www.unb.ca/cic/datasets/nsl.html  
+or use a preprocessed CSV copy.
+
+---
+
+## 📊 Example Outputs
+
+- Model Accuracy Comparison  
+- Confusion Matrices for each model  
+- Bar chart of precision, recall, and F1-score  
+- Sample prediction visualization
+
+---
+
+## 🧰 Technologies Used
+
+- **Python 3.9+**
+- **Streamlit** — web UI
+- **Scikit-learn** — model training
+- **Pandas / NumPy** — data handling
+- **Matplotlib / Seaborn** — visualization
+
 ---
 
-# Welcome to Streamlit!
+## 🏗️ Local Setup (Optional)
+
+```bash
+# Clone the repository
+git clone https://huggingface.co/spaces/your-username/nsl-kdd-anomaly-detection
+cd nsl-kdd-anomaly-detection
 
-Edit `/src/streamlit_app.py` to customize this app to your heart's desire. :heart:
+# Install dependencies
+pip install -r requirements.txt
 
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
+# Run Streamlit app
+streamlit run app.py