A method for early diabetes risk prediction based on deep pca transform

Abstract

The present invention belongs to the field of data processing technology, which involves a method of early diabetic risk prediction based on deep PCA transformation, including the following steps: input early diabetic data sets; data pre -processing, calculating the number of Pilson phase relationships, filtering the redundant characteristics, obtaining the characteristicsEnter data; through the depth PCA extract input data feature collection, as the input of the training logic regression classifier; based on the characteristic set training logic regression classifiers, the judgment of the sample of the case for assessment; enter the new case sample information, output the sampleWhether the results of diabetes are judged and corresponding confidence.The invention realizes the effective extraction of the two -value information of case samples based on the characteristic transformation method based on the deep PCA, and at the same time establish a logical regression classifier to achieve the confidence and quantitative indicator of the confidence of the disease sample and output results, which is convenient and effective to achieve existing existing existing existing existingEarly auxiliary diagnosis of diabetic cases and discovering the condition in time.

Description

technical field

[0001] The invention belongs to the technical field of data processing, and in particular relates to an early stage diabetes risk prediction method based on deep PCA transformation. Background technique

[0002] According to the World Health Organization's 2018 report, diabetes is one of the fastest growing chronic life-threatening diseases, already affecting 422 million people worldwide. Since diabetes has a relatively long asymptomatic period, about 50% of people with diabetes suffer from Prolonged asymptomatic periods go undiagnosed, yet early detection of diabetes is very important for patient management and is only possible through proper evaluation of common and less common signs and symptoms, which can occur between the onset of the disease and the diagnosis of the disease. Different stages are found.

[0003] Reasons for the existence of problems or defects: At present, the diagnosis of diabetes basically relies on clinicopathological analysis, and t...

Images