Abstract
Feature selection (FS) is an important research topic in machine learning. Usually, FS is modelled as a bi-objective optimization problem whose objectives are: 1) classification accuracy; 2) number of features. One of the main issues in real-world applications is missing data. Databases with missing data are likely to be unreliable. Thus, FS performed on a data set missing some data is also unreliable. In order to directly control this issue plaguing the field, we propose in this study a novel modelling of FS: we include reliability as the third objective of the problem. In order to address the modified problem, we propose the application of the non-dominated sorting genetic algorithm-III (NSGA-III). We selected six incomplete data sets from the University of California Irvine (UCI) machine learning repository. We used the mean imputation method to deal with the missing data. In the experiments, k-nearest neighbors (K-NN) is used as the classifier to evaluate the feature subsets. Experimental results show that the proposed three-objective model coupled with NSGA-III efficiently addresses the FS problem for the six data sets included in this study.
Highlights
A large number of data sets contains a lot of irrelevant or redundant features
The “+” denotes that NSGAIII is significantly better than the comparison approach, the “-” denotes that the comparison approach is significantly better than NSGA-III, and “=” denotes that NSGA-III and comparison approach have similar results
This paper proposes a novel interpretation of Feature selection (FS) problem in data science with a specific reference to data sets with missing data
Summary
A large number of data sets contains a lot of irrelevant or redundant features (useless features). After the application of the mean imputation approach, this paper proposes the modelling of the reliability of the data through a third objective of the multi-objective optimization problem. Unlike the studies in the literature, this paper considers the classification accuracy and solution size, and introduces the missing rate for FS in order to enhance upon the reliability of FS. In this study we employ the mean imputation method in single imputation to interpolate the missing data. We chose this method since it is well-suited to handle large data sets thanks to its low computational complexity and modest execution time, see [40]. Where lmj is the number of missing entries associated with the feature j, N is the total number of all instances
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have