Automated-Malware-Analysis

Machine Learning Model to detect hidden malwares and phase changing malwares.It predicts the date of the next probable attack of the malware and its extent.It deals with the change in network traffic flow.It is developed in Python in Jupyter notebook. Developer - Anustup Mukherjee

Malwares are the present date pin-point top notch attacks of cyber crimes to steal data ,spying,hacking the access and all the hustles and bustles going around.Its pattern,signature is changing day by day ,its hiding and polymorphic in nature now a days just like mutating virus.Signatures as well malwares are particularly source of code script that being controleed by the infector sitting over the server and producing continuous injections on anyother server to get the access. They are basically upto calling the APIs or sending the infectious code. It not only infects persons system also steals and spy on Data.Leading companies are facing a data protection problem for this malwares. Malwares are changing there pattern and viral signatures day by day . They are corrupting by embedding themselves in the media files that we transfer now in our social nets also. Darkweb nets are leading malware source to sell and buy these hacked account at a higher rate. This research work is developed by me on the basis of my long work on Malwares at Chandigarh Cyber cell on their data sets of malwares ,crime instances ,real time issues with malware attacks,IIT Patna character and feature analysis of malware attack, Developed product is also presented at Elementor -Microsoft Meet up 2019.The research is went on Microsoft data sets provided by them on malware instance reports.The data sets are made by infecting a particular system by different types of Malware with a specific amount of time. By the repeated infections the nature is studied and developed into customed data set to get used of. The main aim of the research is to detect Hidden and polymorphic malwares ,classify its types, predict the next probable malware attack with the extent of infection and Malware rate of infection monitoring system.The uniqueness of this research is the prediction part as most of the time we are unknown of the extent and forget about protection.Not only this it give promising results on the cases that when people forcefully browse through the Internet inspite of having a warning related to security , it saves in that cases particularly . Antivirus now a days are facing this much challenge to handle this hidden and polymorphic malwares , this research work is based on to solve that issue and change into a fully automated Artificial intelligence Platform leading to new era and approaches of Cyber monitoring. The proposed model is basically a hybrid model approach based on both deep and machine learning approach by making 9 fold trained double neural networks for detection,SVM classifiers for classification of malwares ,Collaboritive filtering for the prediction and monitoring.Hybrid model approaches are always a new way to develop the AI as it promises a mixture of deep and machine learning by using parallel run algorithm techniques with better accuracy and sensitivity of the trained models. The Model is having a accuracy of near about to 95%.The codes are live tested also on several platforms too. Hence the said approach have promising results into the joint field of Cyber security and monitoring with AI. Keywords : SVM,CNN,collaborative filtering,hybrid modelling ,Malwares,DarkWeb, LDA,Net-models,VGG , Alex-net,MAcfree,Kaspersky , Torjan, Spyware , Benign

HOW TO USE THE SCRIPT : DOWNLOAD THE PY SCRIPT : https://github.com/Anustup900/Automated-Malware-Analysis.git 1.) setup your local machine by downloading the Tensor flow backend directly if you are using jupyter motebook or Anaconda !pip install tensorflow !pip install keras 2.) If you are using Google colab the use the command line : docker run -it -p 8888:8888 -p 6006:6006
tensorflow/tensorflow:nightly-py3-jupyter then run : %load_ext tensorboard 3.) And if directly working on downloading python server then start by giving the pip command over the -cmd window

As you are done with the tensor backend then follow with the code script and have a local monitoring by linking malware sites by creating a raw.github link or by calling API or directly calling URL in note book .