Optimal Rank Selection

[mdbarnesUCSD]

Hello,

Does mutSignatures support optimal rank selection?

I am currently seeing three parameters that I can set when using setMutClusterParams().

1. num_processesToExtract
2. num_totIterations
3. num_parallelCores

Is it possible to provide a range for the optimal number signatures to be selected from?

Thanks!

[dami82]

Hello, We are currently evaluating some options to address the problem of optimal k selection. Overall, the following metrics seem very important: cluster stability (silhouette analysis), distances between signatures, and sum of residuals. The consensus at the moment is that an “automatic” selection of k won’t work because datasets and/or biological questions may differ. This being said, I like your idea. It’s not difficult to implement what you suggest for a single argument (num_processes_to_extract). I suppose I should return not just a list of results, but also an object including the aforementioned metrics for each k. Great suggestion, thanks. It may take few weeks before I can address this. Thanks for your patience. Best regards, Damiano —— Damiano Fantini Sent from my iPhone

…

On Nov 5, 2020, at 18:36, mdbarnesUCSD ***@***.***> wrote:  Hello, Does mutSignatures support optimal rank selection? I am currently seeing three parameters that I can set when using setMutClusterParams(). num_processesToExtract num_totIterations num_parallelCores Is it possible to provide a range for the optimal number signatures to be selected from? Thanks! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.