Prediction of intrapartum fetal hypoxia considering feature selection algorithms and machine learning models
Cardiotocography (CTG) consists of two biophysical signals that are fetal heart rate (FHR) and uterine contraction (UC). In this research area, the computerized systems are usually utilized to provide more objective and repeatable results.
Materials and Methods
Feature selection algorithms are of great importance regarding the computerized systems to not only reduce the dimension of feature set but also to reveal the most relevant features without losing too much information. In this paper, three filters and two wrappers feature selection methods and machine learning models, which are artificial neural network (ANN), k-nearest neighbor (kNN), decision tree (DT), and support vector machine (SVM), are evaluated on a high dimensional feature set obtained from an open-access CTU-UHB intrapartum CTG database. The signals are divided into two classes as normal and hypoxic considering umbilical artery pH value (pH < 7.20) measured after delivery. A comprehensive diagnostic feature set forming the features obtained from morphological, linear, nonlinear, time–frequency and image-based time–frequency domains is generated first. Then, combinations of the feature selection algorithms and machine learning models are evaluated to achieve the most effective features as well as high classification performance.
The experimental results show that it is possible to achieve better classification performance using lower dimensional feature set that comprises of more related features, instead of the high-dimensional feature set. The most informative feature subset was generated by considering the frequency of selection of the features by feature selection algorithms. As a result, the most efficient results were produced by selected only 12 relevant features instead of a full feature set consisting of 30 diagnostic indices and SVM model. Sensitivity and specificity were achieved as 77.40% and 93.86%, respectively.
Consequently, the evaluation of multiple feature selection algorithms resulted in achieving the best results.
KeywordsBiomedical signal processing Fetal heart rate Feature selection Classification Machine learning
We would like to thanks to Dr. Sami Güngör due to its comments on the medical background.
There is no funding source for this article.
Compliance with ethical standards
Conflicts of interest
The authors declare that there is no conflict to interest related to this paper.
This article does not contain any data, or other information from studies or experimentation, with the involvement of human or animal subjects.
- 4.Norén H, Amer-Wåhlin I, Hagberg H, Herbst A, Kjellmer I, Marşál K, Olofsson P, Rosén KG. Fetal electrocardiography in labor and neonatal outcome: data from the Swedish randomized controlled trial on intrapartum fetal monitoring. Am J Obstet Gynecol. 2017;188:183–92. https://doi.org/10.1067/mob.2003.109.CrossRefGoogle Scholar
- 11.Ayres-de-Campos D, Ugwumadu A, Banfield P, Lynch P, Amin P, Horwell D, Costa A, Santos C, Bernardes J, Rosen K. A randomised clinical trial of intrapartum fetal monitoring with computer analysis and alerts versus previously available monitoring. BMC Pregnancy Childbirth. 2010;10:71. https://doi.org/10.1186/1471-2393-10-71.CrossRefGoogle Scholar
- 14.Karin J, Hirsch M, Sagiv C, Akselrod S. Fetal autonomic nervous system activity monitoring by spectral analysis of heart rate variations. In: Proceedings of computers in cardiology; 1992. p. 479–482. https://doi.org/10.1109/cic.1992.269517.
- 17.Cömert Z, Kocamaz AF. Cardiotocography analysis based on segmentation-based fractal texture decomposition and extreme learning machine. In: 25th Signal processing and communications applications conference; 2017. p. 1–4. https://doi.org/10.1109/siu.2017.7960397.
- 19.Cömert Z, Kocamaz AF. A study based on gray level co-occurrence matrix and neural network community for determination of hypoxic fetuses. In: International artificial intelligence and data processing symposium, TR, 2016. p. 569–573. https://doi.org/10.13140/rg.2.2.23901.00489.
- 20.Usha Sri A, Malini M, Chandana G. Feature extraction of cardiotocography signal. In: Satapathy SC, Raju KS, Shyamala K, Krishna DR, Favorskaya MN, editors. Advances in decision sciences, image processing, security and computer vision. Cham: Springer; 2020. p. 74–81.Google Scholar
- 24.Uzun A, Kızıltas CE, Yılmaz E. Cardiotocography data set classification with extreme learning machine. In: International conference on advanced technologies, computer engineering and science; 2018. p. 224–230.Google Scholar
- 26.Stylios CD, Georgoulas G, Karvelis P, Spilka J, Chudáček V, Lhotska L. Least squares support vector machines for FHR classification and assessing the pH based categorization. In: Kyriacou E, Christofides S, Pattichis CS, editors. XIV Mediterranean conference on medical and biological engineering and computing, 2016 (MEDICON 2016), 31 March–2 April 2016, Paphos, Cyprus. Cham: Springer; 2016. p. 1211–1215. https://doi.org/10.1007/978-3-319-32703-7_234.CrossRefGoogle Scholar
- 28.Xu L, Georgieva A, Redman CWG, Payne SJ. Feature selection for computerized fetal heart rate analysis using genetic algorithms. In: 2013 35th Annual international conference of the IEEE Engineering in Medicine and Biology Society; 2013. pp. 445–448. https://doi.org/10.1109/embc.2013.6609532.
- 29.Cömert Z, Kocamaz AF. Novel software for comprehensive analysis of cardiotocography signals CTG-OAS. In: Karci A, editor. International conference on artificial intelligence and data processing. Malatya: IEEE; 2017. p. 1–6. https://doi.org/10.1109/idap.2017.8090210.
- 30.Chudacek V, Spilka J, Rubackova B, Koucky M, Georgoulas G, Lhotska L, Stylios C. Evaluation of feature subsets for classification of cardiotocographic recordings. In: 35th Annual conference on computers in cardiology 2008. Piscataway, NJ: IEEE; 2008. p. 845–848. https://doi.org/10.1109/cic.2008.4749174.
- 40.Altuntaş Y, Kocamaz AF, Cömert Z, Cengiz R, Esmeray M. Identification of haploid maize seeds using gray level co-occurrence matrix and machine learning techniques. In: 2018 International conference on artificial intelligence and data Processing, Malatya, Turkey; 2018. p. 1–5. https://doi.org/10.1109/idap.2018.8620740
- 42.Subha V, Murugan D, Boopathi AM, Velappan S, Murugan D, Boopathi M. A hybrid filter-wrapper attribute reduction approach for fetal risk anticipation. Asian J. Res. Soc. Sci. Humanit. 2017;7:1094–106.Google Scholar
- 45.Spilka J, Georgoulas G, Karvelis P, Oikonomou VP, Chudáček V, Stylios C, Lhotská L, Jankru P. Automatic evaluation of FHR recordings from CTU-UHB CTG database. In: Bursa M, Khuri S, Renda ME, editors. Proceedings of 4th international conference on information technology in bio- and medical informatics (ITBAM 2013), Prague, Czech Republic, 28 August 2013. Berlin: Springer, 2013, pp. 47–61. https://doi.org/10.1007/978-3-642-40093-3_4.CrossRefGoogle Scholar