Artificial Intelligence Based Heterogeneous Network For Threat Analysis
Many wireless and wired data nodes are required for the Internet of things (IoT) and 5G networks. To support everything from phone calls to messages or streaming videos to control data transfer, a safe and effective data network is essential for larger Enterprises. The most important application of
2025-06-28 16:25:10 - Adil Khan
Artificial Intelligence Based Heterogeneous Network For Threat Analysis
Project Area of Specialization Cyber SecurityProject SummaryMany wireless and wired data nodes are required for the Internet of things (IoT) and 5G networks. To support everything from phone calls to messages or streaming videos to control data transfer, a safe and effective data network is essential for larger Enterprises. The most important application of the Internet of Things (IoT) has attracted numerous industrial sectors, such as smart cities, heterogeneous networks, etc. because of the evolving industrial revolution. At one point it helps in maintaining business operations and in managing the workflow easily. On the contrary, it poses difficulties for network security. The adoption of the Internet of Things and many other heterogeneous devices has created new issues for network security. As more IoT devices are added, new threats emerge, which the proposed signature-based attack detection systems are unable to detect. Researchers are focusing on current security solutions based on machine learning (ML) algorithms to address these concerns. This project presents the detection of anomalies and cyber-attacks that breaches the system and causes significant damage. The proposed method also prevents cyber-attacks from entering the network by detecting them at the initial phase of the network, generating an alert, and stopping them from causing damage.
Project Objectives- This project aims to conduct threat analysis using intrusion detection for a variety of malware threats which has not been achieved previously.
- By doing this project, gathering a bigger dataset might additionally permit us to peer how assaults detection accuracy is suffering from the quantity and variety of community traffic.
- This project will also help to understand network behavior of different types as it varies widely by device type because certain devices are more amenable to network malware detection.
- Data Capturing: The dataset we are using for this project consists of 220 malware/DDOS samples. The different malware/DDOD samples were acquired by implementing an attack on the systems test folder.
- Data Preprocessing: The data cleansing process has been applied to the dataset so the dataset becomes ready for feature selection. In feature selection, the main feature of the data has been highlighted to detect the object or detect any abnormality of the system. After feature selection the dataset divided into two phases:
- Training Phase (To detect the malware attack train the model on a dataset that consists of different types of malware attack)
- Testing Phase (To test the model to see how much the following model is accurate)
- Detection of malware/DDOS using ANN: The MLP (Multi-layer Perceptron) is an algorithm of ANN which is being used to detect the malware attack. In this system, three layers of the MLP model are being applied. The data travel between the layers with the help of neurons commonly known as Feed Forward Propagation. After data travel forward it will propagate reversely known as Back Propagation.
- Detection of Malware/DDOS Attack: After analyzing the data packets the system separates the malware data and removes it from the network.
- Any malicious activity can be detected quickly.
- This project will improve the security of the network and reduce the threat of malware attacks.
- A report with a brief malware behavior and detection.
The Heterogeneous based network uses an Artificial Intelligence-based Intrusion Detection System to detect anomalies in the networks. The Artificial Intelligence-based Intrusion Detection System is being trained by the different samples of malware and DDOS cyber-attacks.
In real-time deployment, the system captured the data packet through the network analyzer (Scapy). After data capturing the data packet convert into the proper dataset format so that an Artificial Intelligence-based system can understand the dataset. The machine learning algorithm (Multi-Layer Perceptron) Classifies the data packet that either these packets are normal or it is some kind of abnormal (attacks) packet. After classification, the result will be displayed on a monitor using ELK Stack.
SSD are using to increase the processing power of a system because in Artificial Intelligence lot of computing power is being used.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 79500 | |||
| Smart Devices | Equipment | 2 | 800 | 1600 |
| WD Green 480GB SSD | Equipment | 1 | 9900 | 9900 |
| 8GB RAM DDR3 | Equipment | 3 | 6000 | 18000 |
| Switch Cisco Catalyst 3750G-24PS | Equipment | 1 | 40000 | 40000 |
| Documentation (Prints, 4 Books, 4 DvDs) | Miscellaneous | 2 | 5000 | 10000 |