Dynamic enzyme action optimizer for feature selection

Feature selection is a research hotspot in the field of pattern recognition. Due to the nature of high dimensionality and small sample size of transcriptomic data, there exist noise and redundant features. Adopting the feature selection algorithm to reduce the dimension is an important link in omics modeling. In this study, the t-test algorithm was adopted to select important features, and the enzyme action optimizer (EAO) algorithm with a dynamic random search strategy was proposed to search the set of important features. The experimental results prove that the classification performance of the DEAO proposed in this paper in 17 transcriptomic data is superior to the existing feature selection algorithms. This paper proposes a novel and effective feature selection strategy, providing a new method for the analysis of tumors and the recognition of biomarkers.