Malicious URLs detection with autoencoder neural network

This repository contains the source code of Detecting malicious URLs using an autoencoder neural network. An article describing how it works is available at https://www.linkedin.com/pulse/anomaly-detection-autoencoder-neural-network-applied-urls-daboubi/

Requirements

• Python 3.9
• x64 CPU
• Tensorflow-compatible NVIDIA GPU

Install required libraries

pip3 install -r requirements.txt

Merge Inversion blocklist (Google_hostnames.txt) with url_data.csv

python merge_url_data.py

Generated new enriched data

python enrich_urls_data.py

Build and test a model

python train_and_test_urls_autoencoder.py

TODO

To put in place a REST API

Dataset sources

• https://www.kaggle.com/datasets/antonyj453/urldataset
• https://www.kaggle.com/datasets/dfydata/the-online-plain-text-english-dictionary-opted
• https://github.com/elliotwutingfeng/Inversion-DNSBL-Blocklists