Sakni-Tasnim/network-threat-detector

🛡️ Network Threat Detector : Trained Models

Pre-trained models for the AI Network Threat Detector project a full-stack Intrusion Detection System (IDS) trained on the UNSW-NB15 network traffic dataset.

🧠 Models

Model 1 : Binary Classifier (model1_binary.pkl)

Detects whether network traffic is normal or malicious.

Metric	Score
Accuracy	90%
F1 (weighted)	0.90
Training samples	175,341

Model 2 : Multi-Class Classifier (model2_multiclass.pkl)

Identifies the attack type across 9 categories.

Metric	Score
Accuracy	76%
F1 (weighted)	0.77
Classes	DoS, Exploits, Fuzzers, Generic, Reconnaissance, Backdoor, Analysis, Shellcode, Worms

📦 Files

File	Description
model1_binary.pkl	Random Forest binary classifier
model2_multiclass.pkl	Random Forest multi-class classifier
encoders.pkl	Feature encoders
le_attack.pkl	Label encoder for attack types
scaler.pkl	StandardScaler for feature normalization

🗂️ Dataset

UNSW-NB15 University of New South Wales network traffic dataset 175,341 samples | 9 attack categories | Real network flows

🚀 Usage

import joblib
import pandas as pd

# Load models
binary_model = joblib.load('model1_binary.pkl')
multiclass_model = joblib.load('model2_multiclass.pkl')
scaler = joblib.load('scaler.pkl')
encoders = joblib.load('encoders.pkl')

# Scale features
X_scaled = scaler.transform(your_features)

# Step 1 : detect if malicious
is_malware = binary_model.predict(X_scaled)

# Step 2 : classify attack type
attack_type = multiclass_model.predict(X_scaled)

🔗 Full Project

For the complete source code, frontend, and API: 👉 github.com/Sakni-Tasnim/ai-network-threat-detector

👤 Author

Sakni Tasnim Telecommunications & Computer Engineering Student
🔗 GitHub • LinkedIn