Feature Selection Method Based on Differential Correlation Information Entropy

Abstract

Feature selection is one of the major aspects of pattern classification systems. In previous studies, Ding and Peng recognized the importance of feature selection and proposed a minimum redundancy feature selection method to minimize redundant features for sequential selection in microarray gene expression data. However, since the minimum redundancy feature selection method is used mainly to measure the dependency between random variables of mutual information, the results cannot be optimal without consideration of global feature selection. Therefore, based on the framework of minimum redundancy-maximum correlation, this paper introduces entropy to measure global feature selection and proposes a new feature subset evaluation method, differential correlation information entropy. In our function, different bivariate correlation metrics are selected. Then, the feature selection is completed through sequence forward search. Two different classification models are used on eleven standard data sets of the UCI machine learning knowledge base to compare various comparison algorithms, such as mRMR, reliefF and feature selection method with joint maximal information entropy, with our method. The experimental results show that feature selection based on our proposed method is obviously superior to that of other models.

Ideological and Political Teaching Resource Sharing Method Based on Edge Computing

“…Information entropy belongs to the nonsequential index in the ideological and political teaching resource system [20,21]. If the entropy is larger, the disorder of ideological and political teaching resources is significant [22]. Information entropy can judge the uncertainty of ideological and political teaching resources.…”

Classification Feature Optimization for Global Wetlands Mapping

“…The evaluation criteria for the classification capabilities of the features mainly depend on the ability to distinguish between different categories (Dash and Liu, 1997). The separability of features can be divided into several groups, such as entropy function (Wang et al, 2020), Jeffreys-Matusita Distance (JM distance) (Sen et al, 2019), the relative distance between categories (Ell and Ashby, 2012), etc (Sweet, 2003). There have also been studies using Spectral Angular Distance (SAD) and Euclidean distance (ED) to measure the difference between two spectra in different years, which has proven to be the best similarity metric for detecting dual time variation (Huang et al, 2020;Ji et al, 2015).…”

Data‐Driven Detection of Integrated Energy Theft Based on VMD‐MIC and Informer

“…VMD determines the frequency center and bandwidth of each IMF by iteratively searching for the optimal solution of the variational model, thus adaptively realizing the frequency domain segmentation of the signal and IMFs. [ 23 ] Literature [ 24 ] focuses on load decomposition of industrial park load (IPL), an improved variational mode decomposition (IVMD) algorithm that denoises the training data of IPL to improve its stability.…”