Feature weight estimation for gene selection: a local hyperlinear learning approach

Abstract

BackgroundModeling high-dimensional data involving thousands of variables is particularly important for gene expression profiling experiments, nevertheless,it remains a challenging task. One of the challenges is to implement an effective method for selecting a small set of relevant genes, buried in high-dimensional irrelevant noises. RELIEF is a popular and widely used approach for feature selection owing to its low computational cost and high accuracy. However, RELIEF based methods suffer from instability, especially in the presence of noisy and/or high-dimensional outliers.ResultsWe propose an innovative feature weighting algorithm, called LHR, to select informative genes from highly noisy data. LHR is based on RELIEF for feature weighting using classical margin maximization. The key idea of LHR is to estimate the feature weights through local approximation rather than global measurement, which is typically used in existing methods. The weights obtained by our method are very robust in terms of degradation of noisy features, even those with vast dimensions. To demonstrate the performance of our method, extensive experiments involving classification tests have been carried out on both synthetic and real microarray benchmark datasets by combining the proposed technique with standard classifiers, including the support vector machine (SVM), k-nearest neighbor (KNN), hyperplane k-nearest neighbor (HKNN), linear discriminant analysis (LDA) and naive Bayes (NB).ConclusionExperiments on both synthetic and real-world datasets demonstrate the superior performance of the proposed feature selection method combined with supervised learning in three aspects: 1) high classification accuracy, 2) excellent robustness to noise and 3) good stability using to various classification algorithms.

Selection of Significant Clusters of Genes based on Ensemble Clustering and Recursive Cluster Elimination (RCE)

“…SVM-RCE-EC demonstrated improved classification accuracy compared to other methods tested (SVM-RFE [15], SVM-RCE [4], mRMR [16], IMRelief [17], SlimPLS [18] and SMKL-FS [12]) reported by the study of Du et al [12] and this was particularly true for data sets where the 2 classes were difficult to separate and where other methods either failed or had low performance.…”

“…The SVM-RCE-EC approach was compared to 8 previously reported methods [12] including: SVM-RFE [15], SVM-RCE [4], mRMR [16], IMRelief [17], SlimPLS [18] and SMKL-FS [12] using 10-fold CrossValidation (CV) on a variety of datasets. The performance is calculated as the mean effectiveness measurement [12].…”

Diagnostic assessment by dynamic contrast-enhanced and diffusion-weighted magnetic resonance in differentiation of breast lesions under different imaging protocols

“…A feature selection method reported recently, called LHR, uses a highly diagnostic yet compact feature subset [49]. The five features discovered include ADC, Sum Average, Entropy, Elongation and Sum Variance.…”

A novel aggregate gene selection method for microarray data classification

“…The procedure works well on selected genes with higher correlation coefficients based on symmetrical uncertainty. Alternatively, Cai et al [2] initiated a feature weighting algorithm for gene selection called LHR. LHR estimates the feature weights through local approximation based on ReliefF.…”