Detection of fetal distress though a support vector machine based on fetal heart rate parameters

This work aimed at realizing an automatic system for diagnosing fetal sufferance through advanced classification methods applied to reliable indexes extracted from fetal heart rate (FHR) recordings. We selected a set of FHR recordings from a database of 909 exams, which were supplied with the diagnosis at the delivery. The analysis was based on both classical parameters taken from the obstetrical clinical literature and some new indexes already used for HR variability in adults, like the power spectral density (PSD) and the approximate entropy (ApEn). This parameter set was then used as input of a learning machine based on the support vector machine (SVM) algorithm. We obtained a dichotomic classifier, performing the detection of suffering IUGR fetuses from healthy ones. A high percentage of correct classifications, above 84%, was reached by filtering the training set with only 65 of the starting 909 available records