Ensemble of penalized logistic models for classification of high-dimensional data
Musarrat Ijaz, Zahid B Asghar, Asma Gul · Communications in Statistics - Simulation and Computation · 2019
Classification of High-dimensional data, such as gene expression data having far more variables (genes) than observations, is challenging. Classifiers aggregation known as ensemble method has proven improved classification accuracy in a wide range of applications. In this study, we propose an Ensemble of Penalized Logistic models (EPL) which utilizes Ridge regression, Lasso and Elastic net as base learners for ensemble generation. Classification is done on the basis of majority votes of the models. The EPL is assessed on both simulated data and bench mark microarray data sets. Its performance in terms of classification accuracy is compared with state-of-the-art classifiers, i.e., Support Vector Machines (SVM), K-Nearest Neighbors (KNN) and Random Forest (RF). The experimental comparisons show that the EPL has high classification accuracy as compared to the other classifiers considered here.