dotnet / machinelearning-modelbuilder

Simple UI tool to build custom machine learning models.
Creative Commons Attribution 4.0 International
264 stars 56 forks source link

Using PFI for multivariate analysis #1144

Closed ramonsf33 closed 3 years ago

ramonsf33 commented 3 years ago

Is your feature request related to a problem? Please describe. There is no related to a problem, is related to my master degree work and I want to calculate the probability to calculate several features factors

Describe the solution you'd like A clear and concise description of what you want to happen.

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

Additional context Add any other context or screenshots about the feature request here.

LittleLittleCloud commented 3 years ago

hi @ramonsf33 Can you provide more info here.

ramonsf33 commented 3 years ago

Sure

Hi, I want to calculate an index for each breast cancer risk factor using a multivariate analysis. I have a few risk factors and I want to know how much each factor affects the appearance of the disease and to do it numerically, that is, to assign a weight to each risk factor based on the multivariate analysis.

Sent from Mailhttps://go.microsoft.com/fwlink/?LinkId=550986 for Windows 10

From: Xiaoyun Zhangmailto:notifications@github.com Sent: Monday, December 7, 2020 10:20 AM To: dotnet/machinelearning-modelbuildermailto:machinelearning-modelbuilder@noreply.github.com Cc: ramonsf33mailto:ramonsf33@outlook.com; Mentionmailto:mention@noreply.github.com Subject: Re: [dotnet/machinelearning-modelbuilder] New (#1144)

hi @ramonsf33https://github.com/ramonsf33 Can you provide more info here.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/dotnet/machinelearning-modelbuilder/issues/1144#issuecomment-740092764, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AN3GXT5MQTBBHMCAB27GWJDSTUMHJANCNFSM4UPWAPFQ.

JakeRadMSFT commented 3 years ago

Sounds like PFI is what you're looking for.

PFI Docs: https://docs.microsoft.com/en-us/dotnet/machine-learning/how-to-guides/explain-machine-learning-model-permutation-feature-importance-ml-net

I'd take a look at our PFI docs and this issue for code sample: https://github.com/dotnet/machinelearning-modelbuilder/issues/1031

ramonsf33 commented 3 years ago

Hi I have been trying to develop a model using the BCSC database and I can’t be able to do it, could any body help me? I have been trying to load the data from a csv file. The PFI method is the perfect method to me but I need to make it work.

Sent from Mailhttps://go.microsoft.com/fwlink/?LinkId=550986 for Windows 10

From: Jakemailto:notifications@github.com Sent: Monday, December 7, 2020 2:03 PM To: dotnet/machinelearning-modelbuildermailto:machinelearning-modelbuilder@noreply.github.com Cc: ramonsf33mailto:ramonsf33@outlook.com; Mentionmailto:mention@noreply.github.com Subject: Re: [dotnet/machinelearning-modelbuilder] New (#1144)

Sounds like PFI is what you're looking for.

PFI Docs: https://docs.microsoft.com/en-us/dotnet/machine-learning/how-to-guides/explain-machine-learning-model-permutation-feature-importance-ml-net

I'd take a look at our PFI docs and this issue for code sample:

1031https://github.com/dotnet/machinelearning-modelbuilder/issues/1031

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/dotnet/machinelearning-modelbuilder/issues/1144#issuecomment-740207245, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AN3GXTYXEHFMTYZIAPDPK6DSTVGLHANCNFSM4UPWAPFQ.

JakeRadMSFT commented 3 years ago

Can you elaborate on what you've tried and what isn't working?

ramonsf33 commented 3 years ago

Hi sure I have been trying to load the data in on datasets https://www.bcsc-research.org/download_file/view/191/344

and perform a PFI method to obtain the importance of each feature but when I try to load the data using IDataView I got an error saying the column age could not be found.

Sent from Mailhttps://go.microsoft.com/fwlink/?LinkId=550986 for Windows 10

From: Jakemailto:notifications@github.com Sent: Monday, December 14, 2020 10:46 PM To: dotnet/machinelearning-modelbuildermailto:machinelearning-modelbuilder@noreply.github.com Cc: ramonsf33mailto:ramonsf33@outlook.com; Mentionmailto:mention@noreply.github.com Subject: Re: [dotnet/machinelearning-modelbuilder] New (#1144)

Can you elaborate on what you've tried and what isn't working?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/dotnet/machinelearning-modelbuilder/issues/1144#issuecomment-745090600, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AN3GXTYX3LCYOMMUQSSKNATSU4A5TANCNFSM4UPWAPFQ.

LittleLittleCloud commented 3 years ago

Hi @JakeRadMSFT, I'm not quite familiar about PFI API in MLNET. could you help take a look at the column not found error when applying PFI

JakeRadMSFT commented 3 years ago

@ramonsf33 could you share the project or a snippet of the code? I can try to help debug the column issue.

JakeRadMSFT commented 3 years ago

@ramonsf33 Closing the issue for now but I'm happy to help investigate. Feel free to re-open and include the code you've tried.