Metagenome fragment attribute reduction and classification method based on neighbourhood rough set

Abstract

The invention relates to a metagenome fragment attribute reduction and classification method based on a neighbourhood rough set. The method comprises the following steps that: A: randomly obtaining the whole genome sequence of microbial flora, taking the amount and the dimension of each DNA (deoxyribonucleic acid) fragment as condition attributes, and randomly selecting a line and a row as corresponding identification for representing each DNA fragment to serve as a decision attribute so as to form a decision table; B: taking an initial reduction set as an empty set, taking a sample as a whole domain of discourse, calculating the attribute importance degrees of all residual attributes each time, and adding the attribute with the highest attribute importance degree into the reduction set until the attribute importance degrees of all residual attributes are smaller than the lower limit of the set attribute importance degree; and C: classifying the low-dimension genes of the reduced metagenome DNA fragment in the B. Under a situation that the metagenome fragment is not assembled, the metagenome fragment is reduced to obtain classification accuracy which is the same with or higher than classification accuracy before reduction is carried out.

Description

technical field [0001] The invention belongs to the field related to bioinformatics analysis technology, and specifically relates to a method for attribute reduction and classification of metagenomic fragments based on neighborhood rough sets. Background technique [0002] Under the current technical conditions, the number of cultivable microorganisms existing in nature is less than 1% of the total number, that is, the remaining 99% cannot obtain their genetic information through traditional genomics methods. The neighborhood rough set attribute reduction method can overcome the shortcomings of classical rough sets that need to discretize the data when processing continuous data, and can objectively reflect the original appearance of the data. It has been effectively applied in big data analysis, Knowledge dependency discovery, attribute subset selection, decision rule discovery, classification analysis and other fields have important theoretical research value and practical...

AI Technical Summary

Problems solved by technology

[0002] Under the current technical conditions, the number of cultivable microorganisms existing in nature is less than 1% of the total number, that is, the remaining 99% cannot obtain their genetic information through traditional genomics methods

Images