reverseCentaur

Tool to turn bro logs into data for all sorts of machine learning/statisticable csvs

1. AVG/TOT across domains
2. Time series in domain
3. All n length fingerprints for a period across domains
4. All n length fingerprints across domains

To-Do:

Easy

• Add PCR(s)
• Remove grep
• Directory Load
• see if I should be using pandas pivots
• Aggregate each domain file

Hard

• Time Based Split
• Find CDX data
• Data cleaning
• Shitty anomaly detect
• Shitty clustering
• Periodicity & jitter test w/ fake data in iPynb

Long Term:

• https://github.com/featuretools/featuretools
• TPOT
• Random Anomaly Detection