November 2019
Intermediate to advanced
346 pages
9h 36m
English
This is a relatively simple recipe but serves as a good starting point for a more high-powered malicious URL detector. The dataset consists of URLs with the labels 0 and 1, depending on whether they are malicious or benign:
http://google.com,0http://facebook.com,0http://youtube.com,0http://yahoo.com,0http://baidu.com,0http://wikipedia.org,0http://qq.com,0http://linkedin.com,0http://live.com,0http://twitter.com,0http://amazon.com,0http://taobao.com,0http://blogspot.com,0<snip>http://360.cn,0 http://go.com,0 http://bbc.co.uk,0http://xhamster.com,0
In step 1, we train a bidirectional LSTM model. By digging deeper into the code, you can adjust the network to your needs. Having trained our model, it is important to assess its performance ...