Overview

• The goal of the project was to build an Intrusion Detection System (IDS) that can classify the connections in the dataset as an attack or a normal connection
• This was a Binary Classification problem in which we used Fully Connected Neural Networks and Convolutional Neural Networks to classify the target feature
• Broke down the attacks of the target feature into different types and built a model that classified an attack connection to its specific attack type. A Multi-Class Classification used Convolutional Neural Networks was applied to classify the target features.

Methodology

• Applied Data Processing for Data Frame
• One-hot encoded the 9 categorical features of the dataset.
• Normalized all the numeric features of the dataset using z-scores.
• Made all the different types of attacks into one value, attack, to make the problem a Binary Classification problem.
• Used Logistic Regression for Feature Importance Analysis.
• Dropped the 102 features that had a coefficient of between -1 and 1.
• Built a CNN with 9 total layers with two convolutional layers with 32 kernels and 64 kernels and a kernel size of 1 by 5 for multi-class classification.

Result