For the method and system for the feature importance for determining machine learning sample

For the method and system for the feature importance for determining machine learning sample

  • CN 107,316,082 A
  • Filed: 06/15/2017
  • Published: 11/03/2017
  • Est. Priority Date: 06/15/2017
  • Status: Active Application
First Claim
Patent Images

1. a kind of method for being used to determine the feature importance of machine learning sample, including:

  • (A) the essential characteristic subset of machine learning sample is determined, wherein, essential characteristic subset includes at least one essential characteristic;

    (B) importance of machine learning sample multiple target signature subsets to be determined are determined, wherein, each target signatureCollection includes at least one target signature;

    (C) each target signature subset among the multiple target signature subset is directed to, corresponding composite machine is obtainedModel is practised, wherein, the composite machine learning model includes the basic submodel and additional son trained according to lift frameModel, wherein, basic submodel is formed based on essential characteristic trained, and additional submodel is based on each described target signatureTrained is formed;

    And(D) importance of the multiple target signature subset is determined according to the effect of composite machine learning model.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×