Cancer gene classification method and device based on two-stage depth feature selection and storage medium

Abstract

The invention relates to a cancer gene classification method and device based on two-stage depth feature selection and a storage medium. The method comprises the steps: A, training a cancer gene classification model: (1) obtaining training data: in the first stage, integrating three feature selection algorithms to carry out comprehensive feature selection, and obtaining a feature subset; in the second stage, the optimal representation of a feature subset is obtained by using an unsupervised neural network; (2) dividing the optimal representation of the feature subset into a training set and a test set, and inputting the training set and the test set into a neural network for training; and B, cancer gene classification: preprocessing to-be-detected cancer gene data, and inputting the preprocessed to-be-detected cancer gene data into the trained cancer gene classification model to realize cancer gene classification. According to the invention, by using the integrated feature selection method, feature selection is carried out in consideration of all aspects; and the optimal representation of the features is extracted by using the unsupervised neural network, so that cleaner gene features are obtained, and the classification precision is improved.

Description

technical field [0001] The invention relates to a cancer gene classification method, equipment and storage medium based on two-stage deep feature selection, and belongs to the technical field of gene expression. Background technique [0002] Cancer is one of the deadliest diseases in the world. The time of cancer discovery directly determines the treatment effect and life safety of the patient. [0003] The use of machine learning to process gene microarray data sets plays an important role in assisting the early diagnosis of cancer, but the number of gene features in microarray data sets is much larger than the number of samples, resulting in sample imbalance and affecting the efficiency and accuracy of classification. Feature selection is particularly important for gene array data. Existing deep feature selection algorithms are all committed to selecting important features from high dimensions, but they do not consider the large number of retained features and the poor p...

AI Technical Summary

Problems solved by technology

[0013] For the deep feature selection algorithm research so far, many problems have been solved, but there are still problems with a large number of selected features and low classification accuracy. In order to solve these problems, this paper The invention provides a cancer gene classification method based on two-stage deep feature selection, which improves the final classification accuracy through two-stage deep feature selection. The main problems to be solved by the present invention are as follows:

Images