dhyeyinf / ShieldNet

ML-powered Intrusion Detection System trained on ISCXIDS2012 dataset with multi-model comparison, reporting, and automation.

🚨 ISCXIDS2012 - Network Intrusion Detection using Machine Learning

A comprehensive project that benchmarks multiple ML models to detect network intrusions using the ISCXIDS2012 dataset. The system processes day-wise traffic data and evaluates models like Random Forest, SVM, ExtraTrees, Gradient Boost, and more. The pipeline is fully automated and generates performance metrics for all models on all days.

---

📁 Project Structure

ISCXIDS2012-MASTER/
├── algorithms/              # Model implementations
├── data/                    # Dataset (original, CSV, split versions)
│   ├── original/            # Original CSVs from ISCXIDS2012
│   ├── CSV/                 # Cleaned data
│   └── split-CSV/           # 70-30 split data (train-test)
├── plotting/                # Plotting scripts (optional)
├── results/                 # Output metrics and result JSONs
│   ├── single/              # Results from single run per model
│   └── cv/                  # Results from cross-validation
├── runner-scripts/          # Automation scripts
├── ml.py                    # Entry point for training models
├── result_handling.py       # Handles and stores model results
├── extract_all_metrics.py   # Consolidates metrics from results
├── preproc.py               # Preprocessing and data splitting
├── reduction.py             # Feature selection/reduction
├── run_all_model.sh         # Run every ML model on all 6 days
└── README.md                # Project readme

Technologies Used

• Python 3.x
• scikit-learn
• pandas, numpy
• matplotlib (for optional plotting)
• Dataset: ISCXIDS2012

Setup Instruction

1. Clone the Repository

git clone https://github.com/dhyeyinf/ShieldNet.git
cd ShieldNet

2. Install Dependencies Make sure Python is installed (≥3.7), then:

pip install -r requirements.txt

If requirements.txt is missing, you can manually install:

pip install scikit-learn pandas numpy matplotlib

Dataset Structure

The dataset contains six days of network traffic:

-D	Day	Filename	Attack Scenario
0	Monday	TestbedMonJun14Flows.csv	HTTP DoS (Hulk)
1	Tuesday	TestbedTueJun15Flows.csv	DDoS via IRC Botnet
2	Wednesday	TestbedWedJun16Flows.csv	Brute Force
3	Thursday	TestbedThuJun17Flows.csv	SSH Brute Force
4	Saturday	TestbedSatJun19Flows.csv	Additional Brute Force Variants
5	Sunday	TestbedSunJun20Flows.csv	Infiltration Attack

Each file is split into 70% training and 30% testing data during preprocessing.

How to Run

1. Run All Models on All Days From the project root:

chmod +x runner-scripts/run_all_models.sh
./runner-scripts/run_all_models.sh

This will:

• Run every ML model on all 6 days.
• Save JSON results in results/single/<model_name>/

2. Extract Metrics into CSV

python extract_all_metrics.py

This will generate a consolidated CSV report with:

• Accuracy
• Precision
• Recall
• F1-score
• ROC AUC
• Runtime
• Day (-D)
• Model name

Available Models

The following models are evaluated in the project:

• K-Nearest Neighbors (knn)
• Nearest Centroid (ncentroid)
• Decision Tree (dtree)
• Linear SVM (linsvc)
• RBF SVM (rbfsvc)
• Random Forest (rforest)
• AdaBoost (ada)
• Bagging (bag)
• Logistic Regression (binlr)
• Quadratic Discriminant Analysis(qda)
• Linear Discriminant Analysis(lda)
• XGBoost (xgboost)
• Gradient Boost (gradboost)
• Extremely Randomised Trees (extratree)

Each model can also be run individually using:

python ml.py -D 0 -F rforest

Change -D to target day and -F to the model short-name.

Plotting (Optional)

If needed later:

python plotting/plot_single_metrics.py -D 0 -F rforest

Sample Output

After running, the results will be visible in:

results/single/<model_name>/<day>_<model>_Z_<%.json>

And the extracted final CSV report will summarize everything.