Clustering with outliers (DBSCAN)

boefraty — Mon, 12 Dec 2016 15:52:55 GMT

Clustering enables you to find similarity groups in your data, using the well-known density-based spatial clustering of applications with noise (DBSCAN). Unlike many other clustering algorithms, DBSCAN also finds outliers. Settings for the visual let you control and refine algorithm parameters to meet your needs.

Service prerequisites: This R-powered custom visual is used without any modifications in the Power BI service

Desktop prerequisites: To run R scripts in Power BI Desktop, you must separately install R on your local computer.
You can download and install R for free from the Revolution Open download page or the CRAN Repository. For more information, see this article.

R package dependencies (see attached script for int): scales, fpc, car, dbscan

Supports R versions: R 3.3.2, R 3.3.1, R 3.3.0, MRO 3.3.2, MRO 3.3.1, MRO 3.3.0

Re: Clustering with outliers (DBSCAN)

gregor — Mon, 24 Apr 2017 12:55:45 GMT

Thank for your post.

Do you know by any chance how you can identify the outliers?

Meaning how do I find them in my data?

Thank you

Re: Clustering with outliers (DBSCAN)

boefraty — Tue, 16 May 2017 05:57:55 GMT

Hi,

Unfortunatly currently we do not export data from visual. You may need to run R-in-Power Query for that.

topic Clustering with outliers (DBSCAN) in R Script Showcase

Clustering with outliers (DBSCAN)

Re: Clustering with outliers (DBSCAN)

Re: Clustering with outliers (DBSCAN)