The application of hierarchical evolutionary approach for sleep apnea classification

Sleep apnea classification is one principal task that a sleep apnea syndrome automatic diagnostic system should carry out. This paper presents the application of a hierarchical evolutionary algorithm for sleep apnea classification. Without considering the mining methods at each abstraction level, this algorithm provides a unified evolutionary framework to automatically exact knowledge from multivariate time series in real-life applications. It is a hybrid of genetic algorithm and genetic programming, in which several hierarchical levels are expressed with complex hierarchical structures. The preliminary results obtained are discussed.