Human body index-stroke relationship analysis system based on machine learning interpretability

Abstract

The invention discloses a human body index-stroke relationship analysis system based on machine learning interpretability, including a data input module, a data preprocessing module, a machine learning module, a correlation analysis module, an index neighbor search module, and a new sample manufacturing module , The new sample prediction statistics module, through the correlation analysis of the attributes, when an attribute changes, the attributes related to it will also change, so that the generated new samples are closer to the actual situation, so as to study the prevalence of stroke The relationship with the change of the body index, the present invention can obtain the influence of the change of the body index on whether to suffer from stroke, which plays an important role in further research on the prevention of diseases.

Description

technical field [0001] The invention relates to the technical field of data processing and machine learning interpretability, in particular to a model-independent machine learning interpretability method and a disease prediction method. Background technique [0002] With the application and penetration of machine learning algorithms in various fields, the accuracy of the algorithms has been rising, especially the application of deep learning algorithms has further improved the accuracy of machine learning algorithms. However, people cannot fully grasp the working principle of machine learning algorithms, and many algorithms with high enough accuracy are still an unexplainable "black box" for us. [0003] In the medical field, under the background of today's smart medical care, it is of great significance to use machine learning algorithms to predict diseases for "unaffected", which can effectively reduce medical costs. However, in the medical field, the harm of algorithm er...

AI Technical Summary

Problems solved by technology

Therefore, new samples generated by these two methods are sometimes meaningless