adriaanvermaak / malware-detection-using-machine-learning Goto Github PK

View Code? Open in Web Editor NEW

This project analyzes PE information of exe files to detect malware. In this repository you will learn how to create your own dataset and will be able to see the use of machine learning models using the dataset. We will use machine learning for detect malware.

License: MIT License

Jupyter Notebook 100.00%

malware-detection-using-machine-learning's Introduction

Malware-Detection-Using-Machine-Learning

About The Project

This project analyzes PE information of exe files to detect malware.
In this repository you will learn how to create your own dataset and will be able to see the use of machine learning models using the dataset.
We will use machine learning for detect malware.

Getting Started

This pe information was extracted using pefile.

If you want to create your own dataset, check out Data_Set_Generator.ipynb

or you can use data-set/MalwareDataSet.csv

Dataset

This dataset (MalwareDataSet.csv) contains a total of 137,444 data. There are 96,526 safe and 40,918 malware.

It has a total of 9 columns. The information of these columns is as follows:

* AddressOfEntryPoint
* MajorLinkerVersion
* MajorImageVersion
* MajorOperatingSystemVersion
* DllCharacteristics
* SizeOfStackReserve
* NumberOfSections
* ResourceSize
* legitimate

Classification

We used 3 different classification algorithms. These :

* K Nearest Neighbors (KNN)
* Decision Tree
* Random Forest

Results

The Success Rate was calculated as % 97.56118855679404 with the K-Nearest-Neighbors

The Success Rate was calculated as % 98.59433660254359 with the Decision Tree

The Success Rate was calculated as % 99.1240068682518 with the Random Forest

Recommend Projects