rustandruin / spark-skylark-rfm-spectral-clustering Goto Github PK

I want use spectral clustering to predict some samples. but in the code i learn how to input data and make easy predict. but now i find i can't use the data what was normalizered . when i input Normalized data,the values of computerSVD has error. just like, my K equal 8 put into computerSVD then the output U just have one component rather than 8 and the value was same . at last i want know how much is the accuracy when run on your default data. thanks

Hi,

It's great to find your code. I am trying to run your code to implement spectral clustering. But I am not clear why Random Feature Mapping works and what role it plays here. So can you introduce some details about this algorithm? It would be better if you can tell me a link/paper it originates.

Thank you very much!

Regards,
Liang