Affinity propagation clustering-based integrated classifier constructing method

Abstract

The invention relates to an affinity propagation clustering-based integrated classifier constructing method. The method comprises the following steps: S1, preprocessing data; S2, obtaining a characteristic distance negative matrix; S3, clustering characteristics by utilizing an affinity propagation clustering algorithm; S4, constructing base classifiers; S5, repeating the step S4 till the quantity of the base classifiers is up to a preset value; S6, screening the base classifiers; and S7, integrating the base classifiers. The method disclosed by the invention has the advantages that the method can be matched with existing characteristic filtering methods and has a broader application prospect; genes are grouped by utilizing affinity propagation clustering through adopting bicor correlation coefficients as a relevance maxim and characteristic subspaces are constructed in a random selection way on this basis, so that better base classifiers with diversity can be obtained; the base classifiers are fused by utilizing a majority voting method; and therefore, by adopting the method disclosed by the invention, better classifying effect can be obtained and the classifying performance is stable at the same time.

Description

technical field [0001] The invention relates to the fields of bioinformatics and data mining. Especially for gene expression data, an ensemble classifier construction method based on neighbor propagation clustering. Background technique [0002] Cancer, also known as malignant tumor (Malignant neoplasm), is a disease caused by the abnormality of the proliferation mechanism that controls cell growth. In 2011, it surpassed heart disease and became the leading cause of death in the world, and the number of new cases increased every year is increasing. According to the "World Cancer Report 2014" released by the United Nations in February 2014, the number of new cancer cases in 2012 reached 14 million, and by 2030, the number of new cancer cases will increase by 50%, reaching 21.6 million people per year. The outlook for new cancer cases in China is grim. The report pointed out that nearly half of the new cancer cases diagnosed in 2012 occurred in Asia, most of which were in C...

AI Technical Summary

Problems solved by technology

[0011] However, the existing ensemble learning methods are generally unfiltered ensemble learning, and its accuracy needs to be improved. Since it is best to select individuals with better individual effects and large differences between classifier ensembles, all ensembles are not necessarily The optimal solution can be obtained; in addition, the number of base classifiers generated by ordinary clustering methods is fixed, and the differences between individuals are not obvious enough

Images