690698_en-US

liliconcom

Kernel methods to detect various anonymous data

Hi,
As far as I know, JMP does not have any additional anomaly detection except for outlier screening (Explore Outliers).
Some customer wants to develop logic using kernel methods to detect various anonymous data.
In this regard, they asked us  to share information about any cases of kernel methods that have already been developed or similar anomaly detection techniques.
Are there any cases ?
Ikju

Re: Kernel methods to detect various anonymous data

Hi @IkjuShin, that's an interesting use-case. I found this article discussing one simple option for using a kernel as a method for outlier detection. The article shows an approach using Python but these steps could easily be replicated in JMP.

https://bobrupakroy.medium.com/anomaly-detection-kde-8301701ba7b6 

Re: Kernel methods to detect various anonymous data

@IkjuShin if you can share more details about the problem the customer wants to solve and why they want to use anomaly detection, I would like to know more about it.  We have been investigating anomaly detection methods for possible new features in  JMP in the future.