Evaluation of cepstral analysis of EHG signals to prediction of preterm labor

The aim of this paper is to evaluate the application of cepstral analysis for classification of term and preterm labors. We used 20 electrohysterogram records from two groups according to the total length of gestation: term delivery records (pregnancy duration ≥37 weeks) and preterm delivery records (pregnancy duration ≤37 weeks). MLP neural network was employed to classify the two groups. An improved classification accuracy of 72.73% is obtained by using sequential forward feature selection scheme.