Hybrid Particle Swarm Optimization-Fuzzy Inference System for Premature Atrial Contraction Detection

This article presents a new technique to detect a premature atrial contraction (PAC). The technique employs a hybrid of particle swarm optimization (PSO) and fuzzy inference system (FIS), and is called PSO-FIS. In the detection electrocardiographic features are used for the inputs of PSO-FIS. In PSO-FIS, a PSO is used to find the optimal parameters of the FIS. A Gaussian function is employed for the fuzzification part of the FIS. The inputs of the FIS are the interval between two consecutive electrocardiographic R waves and the accumulation of the amplitudes around the P waves. Using clinical data, the technique performs well for PAC detection with 81.93%, 82.27% and 82.26% respectively.