公开(公告)号：US20240193416A1

公开(公告)日：2024-06-13

申请号：US18444906

申请日：2024-02-19

Applicant: SAS Institute Inc.

Inventor： Xinmin Wu ,  Ricky Dee Tharrington, JR. ,  Ralph Walter Abbey ,  Xin Jiang Hunt

IPC: G06N5/022

CPC classification number: G06N5/022

Abstract: A computing device trains a fair prediction model while defining an optimal event cutoff value. (A) A prediction model is trained with observation vectors. (B) The prediction model is executed to define a predicted target variable value and a probability associated with an accuracy of the predicted target variable value. (C) A conditional moments matrix is computed based on fairness constraints, the predicted target variable value, and the sensitive attribute variable value of each observation vector. The predicted target variable value has a predefined target event value only when the probability is greater than a predefined event cutoff value. (D) (A) through (C) are repeated. (E) An updated value is computed for the predefined event cutoff value. (F) (A) through (E) are repeated. An optimal event cutoff value is defined from the predefined event cutoff values used when repeating (A) through (E). The optimal value and prediction model are output.

公开(公告)号：US20230359890A1

公开(公告)日：2023-11-09

申请号：US18208455

申请日：2023-06-12

Applicant: SAS Institute Inc.

Inventor： Xinmin Wu ,  Ricky Dee Tharrington, JR. ,  Ralph Walter Abbey

IPC: G06N20/00

CPC classification number: G06N20/00

Abstract: A computing device trains a fair prediction model. A prediction model is trained and executed with observation vectors. A weight value is computed for each observation vector based on whether the predicted target variable value of a respective observation vector of the plurality of observation vectors has a predefined target event value. An observation vector is relabeled based on the computed weight value. The prediction model is retrained with each observation vector weighted by a respective computed weight value and with the target variable value of any observation vector that was relabeled. The retrained prediction model is executed. A conditional moments matrix is computed. A constraint violation matrix is computed. Computing the weight value through computing the constraint violation matrix is repeated until a stop criterion indicates retraining of the prediction model is complete. The retrained prediction model is output.