Progressive ensemble classification method based on kernel width learning system

Abstract

The invention discloses a progressive ensemble classification method based on a kernel width learning system. The progressive ensemble classification method comprises the following steps: 1) inputtinga training sample and a test sample; 2) training a kernel width learning system as a base classifier by using the original training data; 3) calculating a prediction residual error according to the training result of the first base classifier, and taking the prediction residual error as a label trained by the next base classifier; 4) when the reduction rate of the trained loss function value reaches a threshold value, stopping the training, and not continuing to increase the base classifier any more; and 5) classifying the test samples to obtain a final prediction result. According to the method, the nonlinear fitting capability of the classifier is improved by introducing a kernel mapping technology while redundant back propagation is not needed by utilizing a width learning system, anda plurality of base classifiers are fused by using an integrated means, so that an obvious improvement effect is achieved on a biological information data set with noise, and the accuracy of biological gene classification is improved.

Description

technical field [0001] The invention relates to the technical field of biological gene data analysis, in particular to a progressive integrated classification method based on a kernel width learning system. Background technique [0002] With the popularization and rapid development of mobile Internet, smart medical care came into being. Smart medicine includes the field of research and analysis of biological genes, and machine learning can play a role in mining potential patterns and characteristics in biological information. Data mining of biological information is a difficult task, because biological genes contain many genes that do not express traits and the data noise caused by insufficient collection technology makes some samples not related to genetic diseases in real situations. Matching affects the prediction and classification of biological genes, so a machine learning model that is more robust and resistant to noise is needed. [0003] Although the deep learning ...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art, and propose a progressive integrated classification method based on the kernel width learning system, which can effectively solve the problem that the feature engineering work is too cumbersome, and replace the calculation of the pseudo-inverse by backpropagation The output weight improves the training efficiency, and combines the method of integrated learning to avoid falling into local optimum and overfitting while improving the classification effect of the kernel width learning system, and has good resistance to noise in biological information

Images