Strong variable extracting method and device

Abstract

The invention discloses a strong variable extracting method and device. The method comprises the steps that strong variables are constructed based on fundamental variables in a fundamental variable pool, and a strong variable candidate set used for the k round of iterative training is formed; the k round of iterative training is conducted on a feature selection model on the basis of the strong variable candidate set, wherein the k is a positive integer of which the value is increased successively, and the value of the k is larger than or equal to one; features output by the feature selection model are merged into a selected feature set, and the output features are the strong variables extracted from the strong variable candidate set after the k round of iterative training is conducted on the feature selection model; it is determined that the (k+1) round of iterative training needs to be conducted on the feature selection model on the basis of a residual error of the feature selection model subjected to the k round of iterative training; the fundamental variables used for constructing the features in the selected feature set in the fundamental variable pool are screened out. By means of the strong variable extracting method and device, the strong variables can be effectively extracted from weak variables.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a strong variable extraction method and device. Background technique [0002] Data mining systems in the era of big data often have to face the tasks of screening, processing and modeling a large amount of various types of data. Usually, the direct correlation between most of the data and the target problem or target variable to be solved is weak, and there is also a large amount of data that is redundant with each other, resulting in a situation where there is a large amount of data but little useful information. . [0003] For big data mining systems, how to extract strong variables from many weak variables so as to extract useful information from a large amount of useless information, there is no effective solution in related technologies. Contents of the invention [0004] The present invention provides a strong variable extraction method and device to at least solve th...

AI Technical Summary

Problems solved by technology

Usually, the direct correlation between most of the data and the target problem or target variable to be solved is weak, and there is also a large amount of data that is redundant with each other, resulting in a situation where there is a large amount of data

Images