Abstract
BackgroundThe choice of selection methods to identify important variables for binary classification modeling is critical to produce stable models that are interpretable, that generate accurate predictions and have minimum bias. This work is motivated by data on clinical and laboratory features of severe dengue infections (dengue hemorrhagic fever, DHF) obtained from 51 individuals enrolled in a prospective observational study of acute human dengue infections.ResultsWe carry out a comprehensive performance comparison using several classification models for DHF over the dengue data set. We compared variable selection results by Multivariate Adaptive Regression Splines, Learning Ensemble, Random Forest, Bayesian Moving Averaging, Stochastic Search Variable Selection, and Generalized Regularized Logistics Regression. Model averaging methods (bagging, boosting and ensemble learners) have higher accuracy, but the generalized regularized regression model has the highest predictive power because the linearity assumptions of candidate predictors are strongly satisfied via deviance chi-square testing procedures. Bootstrapping applications for evaluating predictive regression coefficients in regularized regression model are performed.ConclusionsFeature reduction methods introduce inherent biases and therefore are data-type dependent. We propose that these limitations can be overcome using an exhaustive approach for searching feature space. Using this approach, our results suggest that IL-10, platelet and lymphocyte counts are the major features for predicting dengue DHF on the basis of blood chemistries and cytokine measurements.
Highlights
The choice of selection methods to identify important variables for binary classification modeling is critical to produce stable models that are interpretable, that generate accurate predictions and have minimum bias
This happens because without prior removal of irrelevant genes, the classification problem is known as the small sample size problem
We addressed the multiple testing using false discovery rate (FDR), a resampling method have an improved ability to reject a false hypothesis compared to the Bonferroni correction
Summary
The choice of selection methods to identify important variables for binary classification modeling is critical to produce stable models that are interpretable, that generate accurate predictions and have minimum bias. Analysis of high-dimensional data, where the number of predictors exceeds the sample size, poses many challenges for statisticians and calls for new statistical methodologies in order to select relevant variables that are correlated with one-another. In these data, feature selection is used to overcome the curse of dimensionality by removing non-essential variables to achieve a model with predictive accuracy. Various blood chemistries and cytokine measurements can be associated with disease outcome Many of these variables are highly correlated and which of the factors will result in the most stable classifier is not known.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
