Microarray cancer gene expression datasets are high dimensional and thus complex for efficient computational analysis. In this study, we address the problem of simultaneous gene selection and robust classification of cancerous samples by presenting two hybrid algorithms, namely Discrete firefly based Support Vector Machines (DFA-SVM) and DFA-Random Forests (DFA-RF) with weighted gene ranking as heuristics. The performances of the algorithms are then tested using two cancer gene expression datasets retrieved from the Kent Ridge Biomedical Dataset Repository. Our results show that both DFA-SVM and DFA-RF can help in extracting more informative genes aiding to building high performance prediction models. © 2013 Springer.