sobanahmed6061/ddos-mitigation-models

DDoS Mitigation Tool — ML Models

Pre-trained models for DDoS attack detection and classification.

Models Included

File	Model	Purpose	Performance
isolation_forest.joblib	Isolation Forest	Anomaly detection	F1: 0.9485
lstm_model.keras	LSTM Neural Network	Temporal analysis	F1: 0.9804
classifier_rf.joblib	Random Forest	Attack classification	Acc: 100%
scaler.joblib	StandardScaler	IF feature scaling	—
lstm_scaler.joblib	MinMaxScaler	LSTM feature scaling	—
classifier_scaler.joblib	StandardScaler	RF feature scaling	—
classifier_encoder.joblib	LabelEncoder	Class encoding	—
model_meta.json	Metadata	IF model info	—
lstm_meta.json	Metadata	LSTM model info	—
classifier_meta.json	Metadata	RF model info	—

Attack Types Detected

• SYN Flood
• UDP Flood
• ICMP Flood
• HTTP Flood
• Slowloris
• Volumetric

Training Data

• Normal traffic: 169 windows
• Attack traffic: 133 windows (5 attack types)
• Window size: 5 seconds
• Features: 20 per window

Usage

import joblib
import numpy as np

# Load models
iso_model  = joblib.load('isolation_forest.joblib')
iso_scaler = joblib.load('scaler.joblib')

# Predict (1=normal, -1=anomaly)
features = np.array([[...]])  # 20 features
x_scaled = iso_scaler.transform(features)
prediction = iso_model.predict(x_scaled)

Full Project

GitHub: https://github.com/sobanahmed6061/ddos-mitigation-tool