Multi-objective optimization-based high-dimensional data semi-supervised ensemble classification method

Abstract

The invention discloses a multi-objective optimization-based high-dimensional data semi-supervised ensemble classification method, relates to the field of artificial intelligent ensemble learning, and mainly solves the problems in sub-space optimization and selection and semi-supervised information utilization in the prior art. The high-dimensional data semi-supervised ensemble classification method comprises the following steps of S1, inputting a training data set; S2, performing data normalization on the input training data set; S3, generating a random sub-space set; S4, performing multi-objective optimization selection of the sub-space set; S5, searching for an optimal semi-supervised classifier in feature sub-spaces; S6, classifying test samples; and S7, calculating classification accuracy. According to the method, the problem of difficulty in processing high-dimensional data is solved by utilizing random sub-spaces; the selection of the sub-spaces is fully optimized by utilizing a multi-objective optimization solution to improve the robustness of the sub-spaces; and the classifiers of the sub-spaces are optimized by fully utilizing untagged and tagged information, so that the generalization capability of the classifiers is improved.

Description

technical field [0001] The invention relates to the field of computer artificial intelligence, in particular to a semi-supervised integrated classification method for high-dimensional data based on multi-objective optimization. Background technique [0002] The acquisition of labeled data requires a lot of manpower and material resources, but unlabeled data is often easier to collect. For example, there is a large amount of unmarked webpage information in text mining, and it would be unimaginable to spend time and other expenses if manually marking the webpage information one by one. Since the classification effect of the learning model trained by supervised learning will be relatively poor when the number of training samples is insufficient, and in order to solve this problem, more and more researchers have focused their attention on how to use unlabeled and In the field of semi-supervised classification of labeled data, such as video annotation, image annotation, natural ...

AI Technical Summary

Problems solved by technology

Third, some methods that use random subspaces simply sample attributes without optimizing and selecting each subspace

Fourth, in the integrated learning method, the results of each subspace are generally combined to obtain the final result, but the current method is only a simple voting method, and cannot classify each selected subspace. evaluate the process accordingly

Images