McTwo: a two-step feature selection algorithm based on maximal information coefficient

Abstract

BackgroundHigh-throughput bio-OMIC technologies are producing high-dimension data from bio-samples at an ever increasing rate, whereas the training sample number in a traditional experiment remains small due to various difficulties. This “large p, small n” paradigm in the area of biomedical “big data” may be at least partly solved by feature selection algorithms, which select only features significantly associated with phenotypes. Feature selection is an NP-hard problem. Due to the exponentially increased time requirement for finding the globally optimal solution, all the existing feature selection algorithms employ heuristic rules to find locally optimal solutions, and their solutions achieve different performances on different datasets.ResultsThis work describes a feature selection algorithm based on a recently published correlation measurement, Maximal Information Coefficient (MIC). The proposed algorithm, McTwo, aims to select features associated with phenotypes, independently of each other, and achieving high classification performance of the nearest neighbor algorithm. Based on the comparative study of 17 datasets, McTwo performs about as well as or better than existing algorithms, with significantly reduced numbers of selected features. The features selected by McTwo also appear to have particular biomedical relevance to the phenotypes from the literature.ConclusionMcTwo selects a feature subset with very good classification performance, as well as a small feature number. So McTwo may represent a complementary feature selection algorithm for the high-dimensional biomedical datasets.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-016-0990-0) contains supplementary material, which is available to authorized users.

Correlated and weakly correlated fault detection based on variable division and ICA

“…The MIC with a range ½0; 1 is normalised and symmetric. Moreover, a high value of MIC indicates a high correlation between the corresponding variables, whereas MIC ¼ 0 indicates that two corresponding variables are independent (Ge et al, 2016). The MIC of two variables x i and x j is defined as follows (Zhang, Jia, Huang, Qiu, & Zhou, 2014):…”

Machine learning based detection of age-related macular degeneration (AMD) and diabetic macular edema (DME) from optical coherence tomography (OCT) images

“…Wrappers screen for a subset of features with an optimal performance measurement, which is usually the classification accuracy or error rate. Wrappers are usually slower but more accurate than filters [32]. So this study chose two wrapper feature selection algorithms CFS (Correlation-based Feature Subset) and CSE (Classifier Subset Evaluator) to find the feature subset significantly associated with the phenotypes [33,34].…”

“…CSE reduced the numbers of features from 493 to only 23 features. Previous comparative studies demonstrated that the above feature selection algorithms usually performed best on selection a feature subset with good classification performances, while feature selection algorithms based on the method of filter like t-test optimize the phenotype association significances [32].…”

“…Classification algorithms are also essential to train a good model [32,37]. We tested our data set using one representative from each classification algorithm group, i.e.…”

Relevance popularity: A term event model based feature selection scheme for text classification

“…The method and software tool got good performance on several bioinformatics problems [14–16]. Zhou’s lab (Health Informatics Laboratory) described a feature selection algorithm, McTwo, to select features associated with phenotypes, independently of each other, and achieving high classification performance [17]. While, unsupervised methods select features when the document class labels are absente [18–20].…”