Feature selection method based on attribute condition redundancy

Abstract

The invention discloses a feature selection method based on attribute condition redundancy. The method comprises the following steps: step 1, preprocessing a data set; 2, dividing the preprocessed data set into a training data set and a test data set; 3, calculating a mutual information value between each target feature in the training data set and a class label, selecting the target feature which enables the current mutual information value to be maximum, deleting the target feature from the original data set, adding the target feature into a set S which is initially empty, and then iteratively performing feature selection according to a feature selection algorithm based on attribute condition redundancy; and adding the features selected by each iteration into the set S, and carrying out iteration to finally obtain a feature subset with the size of m. According to the method, redundant information of two measurement features, namely redundancy and attribute condition redundancy, is used, so that some redundant features are removed more accurately, and the classification correctness of the test data is improved.

Description

technical field [0001] The invention belongs to the technical field of data mining and relates to a feature selection method based on attribute condition redundancy. Background technique [0002] Due to the rapid development of computer technology and data storage capacity, data processing capabilities have made a qualitative leap, and various big data are used in scientific research and social life. However, the continuous increase of data dimensions makes data processing more and more difficult. It is extremely urgent and important to study data dimensionality reduction technology. The data after dimensionality reduction not only effectively reduces the feature dimension, but also reduces the space required for the feature data storage, and also reduces the difficulty of obtaining key information between the main features, because the key features are retained, and the subsequent The time cost of algorithmic models searching for key features. Dimensionality reduction te...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a feature selection method based on attribute condition redundancy, which solves the problem that the three-way feature interaction cannot be measured in the existing feature selection framework, and the selected feature subset contains more redundant information.

Images