Method for predicting methylene blue treatment response of infectious shock patient based on interpretable model

The invention discloses a method for predicting methylene blue treatment response of an infectious shock patient based on an interpretable model. The method comprises the following steps: firstly, constructing a methylene blue treatment response prediction system of the infectious shock patient, which consists of a data preprocessing module, a feature engineering module and a response prediction module; the feature engineering module carries out feature engineering on the data set and constructs two groups of training sets and two groups of test sets. A model training sub-module in the response prediction module adopts two groups of training sets to train five machine learning models respectively to obtain ten response prediction models. And the model screening sub-module evaluates the response prediction model by using the two test sets, and screens out a final response prediction model. And the model interpretation sub-module interprets the final response prediction model to obtain a final prediction model,