Unsupervised clustering for fetal state assessment based on selected features of the cardiotocographic signals

In modern obstetrics the cardiotocography is a rout ine method of fetal condition assessment based main ly on analysis of the fetal heart rate signals. The correct interpretatio n of recorded traces from a bedside monitor is very difficult even for experienced clinicians. Therefore, computerized fetal monitorin g systems are used to yield the quantitative descri ption of the signal. However, the effective techniques enabling automated conclus ion generation based on cardiotocograms are still b e ng searched. The paper presents an attempt to diagnose the fetal state bas ing on seventeen features describing the cardiotoco graphic records. The proposed method applies the unsupervised classification of s ignals. During our research we tried to classify th e fetal state using the fuzzy c-means (FCM) clustering. We also tested how the eff ici ncy of classification could be influenced by ap plication of principal component analysis (PCA) algorithm. The obtained res ults howed that unsupervised classification cannot be considered as a support to fetal state assessment.