DBSCAN

Outlier detector based on DBSCAN.

Fairly unreliable as it is very sensitive to input parameters which make it difficult to automatically determine suitable parameters. Automatically determined parameters can sometimes be very effective, but often they do not have the desired results. Seeing as there is a single epsilon value for all clusters the algorithm fails when varying density clusters are present in the data.

Therefore if DBSCAN identifies more than 33% of the data points in a time series as outliers, this algorithm will return an inconclusive results.

See the docstrings - https://earthgecko-skyline.readthedocs.io/en/latest/skyline.custom_algorithms.html#module-custom_algorithms.dbscan

See the custom_algorithm source - https://github.com/earthgecko/skyline/blob/master/skyline/custom_algorithms/dbscan.py