I am new to the world of ML and data mining and looking for the help and guidance to find unusual behavior on my log file.
Assuming I have a cvs file which logs users sessions start time and end time and the policy numbers that they have worked on similar to below.
Start_date, username, end_date, Policy_numbers
2018-01-02 10:01, user1, 2018-01-02 10:10, PO-123
2018-01-02 10:05, user2, 2018-01-02 10:20, PO-456
2018-01-02 10:11, user1, 2018-01-02 10:45, PO-789 | PO-999 (| is delimiter here)
Is there any Python or java library/module/code or open-source application to identify patterns such as : Most users logged in during 10 AM to 5 PM , Average number of sessions per days in a month, Average length of session and … , I expect application recognize various of patterns and suggest it to me in a list or way that I can pick those one which matters to business.
(If I recognize the pattern then I can find the answers by some queries and no need for pattern recognition- that would be an easy job to do)
Then is there a way to train the system by these recognized patterns to find unusual behaviors such as : find users who logged in after 5:00 PM , find sessions took way longer than average and ...
Thanks for any thought.
