Spam user detection report
Download as PPTX, PDF2 likes321 views
The document discusses spam user detection through machine learning tools from AWS and Azure. It proposes two algorithms: detecting spam users by their time frequency of actions, flagging those with over 1 million actions; and detecting spam users by analyzing correlations between the articles they interact with, flagging those with high correlations between article titles and counts. The presentation thanks the audience at the end.
Spam user detection report
- 1. Spam User Detection Andy, Len, Petertc
- 2. Agenda Machine Learning Tools AWS Machine Learning Azure Machine Learning Spam Detection Algorithm Detect spam user by time frequency. Detect spam user by article correlation.
- 6. Azure Machine Learning https://manage.windowsazure.com
- 8. Detect spam user by time frequency.
- 9. Detect spam user by time frequency. 500,000 actions
- 10. Detect spam user by time frequency. 1,000,000 actions 500,000 actions
- 11. Detect spam user by time frequency. 1,000,000 actions 3,000,000 actions
- 12. Detect spam user by article correlation. Spam User Spam User Spam User High correlation High correlation High correlation
- 13. Detect spam user by article correlation. Title_count Article_count High correlation
- 14. Thanks!