Classification model output calibration method and device and readable storage medium

The invention relates to an output calibration method and device of a classification model and a readable storage medium, and belongs to the technical field of computers. Comprising the following steps: constructing a calibration model based on a principle of classification decision invariance after output to probability mapping; obtaining a training set, dividing the training set into a plurality of training subsets, inputting the training subsets into the classification model, and calculating the average accuracy of the classification model based on a classification result; inputting the classification result into the calibration model, and calculating the average confidence of the classification model; and constructing an optimization function by taking the minimum mean square expectation error of the average accuracy and the average confidence coefficient of the classification model as a target, adjusting parameters of the calibration model by using a stochastic gradient descent method, and continuously i