Raman spectrum data classification method based on self-encoding network

Abstract

The invention provides a Raman spectrum data classification method based on a self-encoding network. The method comprises the steps of extracting a Raman spectrogram of platelets of patients sufferingfrom Alzheimer's disease in different disease ages through an optical tweezer Raman system, and randomly dividing the pretreated Raman spectrogram into a training set and a testing set; constructinga stack-type sparse self-encoding network with a two-layer feature layer through a layer-by-layer greedy training method by taking a training set sample as input data; training a Softmax classifier with a deep feature of the second-layer autocoder as the input data, and replacing an output layer of the stack-type sparse self-encoding network with the two-layer feature layer by taking the trained Softmax classifier as a Softmax classification layer; constructing an intial classification network according to the stack-type sparse self-encoding network with the two-layer feature layer, optimizingthrough a back propagation algorithm to obtain an optimized classification network; and obtaining a nerve cell quantity optimized classification network through optimization of a nerve cell quantityby taking the testing set as the input data of the optimized classification network. With the Raman spectrum data classification method based on the self-encoding network, the classification accuracyand stability are improved.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence applications, and in particular relates to a Raman spectral data classification method based on an autoencoder network. Background technique [0002] Alzheimer's Disease (Alzheimer's Disease) is the most common disease type in senile dementia. It is a degenerative disease of the central nervous system and a chronic degenerative disease that can cause damage to the brain. The clinical symptoms are cognitive dysfunction, memory impairment, and language impairment, which directly affect the normal life of patients. According to the data in 2006, there are more than 200 million AD patients in the world. Moreover, the degree of aging society is getting higher and higher, and there are more and more AD patients. Alzheimer's disease has become a disease that the whole society has to face. Each AD patient will bring a heavy burden to a family or several families and indirectly affect all...

AI Technical Summary

Problems solved by technology

But there is usually such a situation that when some samples in the input training samples are quite different from other samples, the decision boundary of the classification will be greatly affected

Images