LOSS FUNCTION OPTIMIZATION METHOD AND APPARATUS, COMPUTER DEVICE, AND STORAGE MEDIUM

摘要

Disclosed in the present application are a loss function optimization method and apparatus, a computer device, and a storage medium. The method comprises: if initial samples are initial unbalanced samples, obtaining a positive sample ratio and a negative sample ratio in the initial unbalanced samples; obtaining a prediction model obtained by corresponding training according to the initial unbalanced samples, and obtaining a predicted value corresponding to each sample according to the prediction model; obtaining a value interval satisfied by the predicted value and a true value of each sample in the initial unbalanced samples, and according to the value interval satisfied by the predicted value and the true value of each sample and a preset adjustment value setting strategy, obtaining a loss function adjustment value corresponding to each sample in the initial unbalanced samples; and multiplying the loss function adjustment value corresponding to each sample in the initial unbalanced samples by a cross entropy of the corresponding sample, and summing to obtain a loss function corresponding to the prediction model.