Clinical prediction model decision analysis method

Abstract

The invention relates to a clinical prediction model decision analysis method. The method comprises the following steps: for a sample data set consistent with a data structure type in a target table,training by applying a preset machine learning algorithm in combination with a binary type target of each object in the sample data set to obtain a target prediction model corresponding to the targettable; evaluating an effect value pb corresponding to each division value in a range of 0 to 1; and selecting an optimal threshold value based on the size of each effect value pb to serve as a boundary value of the probability value range of the object in the target table based on the dichotomy type target corresponding to the target prediction model, thereby further targeting each object in the target table. According to the data of each attribute corresponding to the object in the target table, the judgment of the preset dichotomy type target corresponding to the object is realized, the working efficiency of dichotomy type target judgment based on multi-dimensional attributes can be greatly improved, and the scheme has very high application value in clinical research.

Description

technical field [0001] The invention relates to a clinical prediction model decision analysis method, which belongs to the technical field of clinical medical statistics. Background technique [0002] Now the clinical prediction model is an important tool for auxiliary analysis of many diseases, which can help clinical personnel to quickly summarize and comprehensively compare various information of patients, and make treatment strategies. A reliable and effective clinical prediction model is established based on a large amount of clinical data, which can be included in the comprehensive analysis of multiple clinical variables, combined with medical statistics or machine learning methods, to obtain the probability of a certain event occurring in a certain sample. According to the probability value, it is considered whether to give the patient a certain method of treatment, and it is expected that the treatment method can achieve the best and least damaging effect. After the...

AI Technical Summary

Problems solved by technology

Although many methods have been used to determine the optimal threshold, the threshold found based on theoretical methods may not necessarily bring the best results for the clinical prediction model obtained in specific experiments. When considering a single threshold, it is more important to first analyze certain The size of the threshold effect within the range and how to evaluate the effect of different range thresholds is one of the difficulties in the follow-up analysis of clinical prediction models

Images