How to Apply the Machine Learning Appropriate Way for Your Security Operation
summary
S4 integrates 17 tool scores into three similarity dimensions and replaces brittle piecewise linear tuning with logistic regression to auto-tune weights so qualitative same/different labels become a usable numeric priority; multicollinearity is handled with PCA and resampling (over/under‑sampling) and overlapping distributions are excluded by simple thresholds; evaluation uses seven families drawn from VirusTotal and CIS monthly reports, yields clustering that connects most same‑family samples and matches a commercial service’s accuracy.
AI-written summary · about
chapters (19)
- 0:00Introduction
- 0:27About Me
- 1:32Security Operation Field
- 2:12Manage Security Operation
- 3:12Structural System
- 4:25Problem
- 5:53Planning
- 7:51Data Categories
- 8:28Quantitative Data
- 9:28Multivariable Analysis
- 10:01Logistic Regression
- 11:30Coding
- 12:36Trial Errors
- 16:03Inbalanced Data
- 20:24Verification
- 21:43Evaluation
- 22:06Results
- 23:22Qualitative Evaluation
- 24:40Take Away
related talks
32mAI/Machine Learning: What is Actually Working in Cybersecurity?
30mData Driven Security
40mDemystifying File Similarity for Malware Detection
01h 01mD1T2 - Applying Machine Learning to User Behavior Anomaly Analysis
30mMachine Learning Techniques for Class Imbalances & Adversaries
22mMachine Learning Based Attack Vector Modeling for CyberSecurity
40mML with Domain-Specific Ontology for IT Security Industry
39mDetecting Website Intrusion and Account Compromise with Machine Learning