Hybrid Filter-wrapper Feature Selection Using Water Wave Optimization for Financial Crisis Prediction in Enterprises

Establishing an effective financial crisis prediction model has profound significance to enterprises’ survival and development. In order to develop accurate financial crisis prediction model, the key step is to select some appropriate features or variables relevant to the prediction, which is called feature selection problem. Appropriate features can contribute greatly to the prediction performance. In this paper, in order to select appropriate features, we propose a hybrid filter-wrapper approach. In the filter phase, we employ the information gain and Spearman correlation as evaluation criteria. In the wrapper phase, we integrate the water wave optimization (WWO), which is adapted based on the solution space of feature selection problem. The computational experiments demonstrate that the adapted algorithm performance better than the compared algorithms on the test instances.