A Smart System for Detecting Behavioural Botnet Attacks using Random Forest Classifier with Principal Component Analysis

Over the years, malware (malicious software) has become a major challenge for computer users, organizations, and even countries. In particular, a compromise of a set of inflamed hosts (aka zombies or bots) is one of the severe threats to Internet security. Botnet is described as some computer systems or devices controlled on the Internet to carry out unintentional and malicious acts without the owner's permission. Due to the continuously progressing behavior of botnets, the conventional methods fail to identify botnets. In other to solve the stated problem, this paper presents a smart system for detecting behavioural bootnet attacks using Random Forest Classifier and Principal Component Analysis (PCA). The system starts with a botnet dataset that was used in building a robust model in detecting Bootnet attacks. The dataset was pre-processed using pandas library for data cleaning. PCA was used in reducing the dimension of the dataset, so as to avoid data imbalance. The result of the PCA was used as input to the random forest classifier. The random forest classifier was trained using the number of estimators as 1000. The result of the model shows a promising accuracy of about 99%.