This is an anomaly detector that finds transactions with unexpected items co-ocurring together. This is in contrast with most anomaly detection schemes which look at transactions more holistically. This is a scikit-learn-compatible, partial implementation of Bertens, Roel, Vreeken, Jiles and Siebes, Arno (2017), Efficently Discovering Unexpected Pattern Co-Occurences, SDM17.
jdyer1/UnexpectedPatternCoocurrence
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|