rcf3 0.3.0

A Rust implementation of the Random Cut Forest algorithm for anomaly detection.
Documentation
1
2
3
4
5
6
7
8
9
10
11
# RCF3 documentation

`rcf3` exposes two public detector families:

- [Forest API]forest.md: Random Cut Forest for numerical streaming data, including anomaly scores, attribution, neighborhood search, imputation, forecasting, and serialization
- [MStream API]mstream.md: mixed numerical/categorical streaming anomaly detection with logical timestamps and decomposed scores

Start with the guide that matches the shape of your data:

- choose **Forest** for numerical observations and the full RCF feature set
- choose **MStream** when events combine numerical and categorical aspects