A Study of Artificial Neural Network Training Algorithms for Classification of Cardiotocography Signals

Year 2017, , 93 - 103, 26.12.2017

Zafer Cömert , Adnan Kocamaz

https://doi.org/10.17678/beuscitech.338085

Cited By: 47

Abstract

Cardiotocography (CTG) that
contains fetal heart rate (FHR) and uterine contraction (UC) signals is a
monitoring technique. During the last decades, FHR signals have been classified
as normal, suspicious, and pathological using machine learning techniques. As a
classifier, artificial neural network (ANN) is notable due to its powerful
capabilities. For this reason, behaviors and performances of neural network
training algorithms were investigated and compared on classification task of
the CTG traces in this study. Training algorithms of neural network were
categorized in five group as Gradient Descent, Resilient Backpropagation,
Conjugate Gradient, Quasi-Newton, and Levenberg-Marquardt. Two different experimental
setups were performed during the training and test stages to achieve more
generalized results. Furthermore, several evaluation parameters, such as accuracy
(ACC), sensitivity (Se), specificity (Sp), and geometric mean (GM), were taken
into account during performance comparison of the algorithms. An open access CTG
dataset containing 2126 instances with 21 features and located under UCI
Machine Learning Repository was used in this study. According to results of
this study, all training algorithms produced rather satisfactory results. In
addition, the best classification performances were obtained with
Levenberg-Marquardt backpropagation (LM) and Resilient Backpropagation (RP) algorithms.
The GM values of RP and LM were obtained as 89.69% and 86.14%, respectively. Consequently,
this study confirms that ANN is a useful machine learning tool to classify FHR
recordings.

Keywords

Biomedical Signal Processing, Fetal Heart Rate, Artificial Neural Network, Training Algorithm, Classification

References

• Amato F, López A, Peña-Méndez EM, et al. (2013) Artificial neural networks in medical diagnosis. Journal of Applied Biomedicine 11(2): 47–58. Available from: http://www.sciencedirect.com/science/article/pii/S1214021X14600570.
• Ayres-de-campos D, Bernardes J, Garrido A, et al. (2000) SisPorto 2.0: A Program for Automated Analysis of Cardiotocograms. Journal of Maternal-Fetal Medicine 9(5): 311–318. Available from: http://www.tandfonline.com/doi/abs/10.3109/14767050009053454.
• Ayres-de-Campos D, Spong CY and Chandraharan E (2015) FIGO consensus guidelines on intrapartum fetal monitoring: Cardiotocography. International Journal of Gynecology & Obstetrics 131(1): 13–24.
• Azar AT (2013) Fast neural network learning algorithms for medical applications. Neural Computing and Applications, Springer 23(3–4): 1019–1034.
• Basheer IA and Hajmeer M (2000) Artificial neural networks: fundamentals, computing, design, and application. Journal of Microbiological Methods 43(1): 3–31. Available from: http://www.sciencedirect.com/science/article/pii/S0167701200002013.
• Battiti R (1992) First-and second-order methods for learning: between steepest descent and Newton’s method. Neural computation, MIT Press 4(2): 141–166.
• Cesarelli M, Romano M, Bifulco P, et al. (2007) An algorithm for the recovery of fetal heart rate series from CTG data. Computers in Biology and Medicine 37(5): 663–669.
• Chudacek V, Spilka J, Rubackova B, et al. (2008) Evaluation of feature subsets for classification of cardiotocographic recordings. In: Computers in Cardiology (CinC), pp. 845–848. Available from: http://ieeexplore.ieee.org/ielx5/4729059/4748952/04749174.pdf?tp=&arnumber=4749174&isnumber=4748952.
• Cömert Z and Kocamaz AF (2017) Cardiotocography analysis based on segmentation-based fractal texture decomposition and extreme learning machine. In: 25th Signal Processing and Communications Applications Conference (SIU), pp. 1–4.
• Cömert Zafer and Kocamaz AF (2017) Comparison of Machine Learning Techniques for Fetal Heart Rate Classification. Acta Physica Polonica A. Cömert Z, Kocamaz AF and Gungor S (2016) Cardiotocography signals with artificial neural network and extreme learning machine. In: 24th Signal Processing and Communication Application Conference (SIU).
• Czabanski R, Jezewski J, Matonia A, et al. (2012) Computerized analysis of fetal heart rate signals as the predictor of neonatal acidemia. Expert Systems with Applications 39(15): 11846–11860.
• Demuth H, Beale M and Hagan M (2010) Neural Network Toolbox TM 6 User’s Guide. Network, The MathWorks, Inc.
• Dennis Jr JE and Schnabel RB (1996) Numerical methods for unconstrained optimization and nonlinear equations. SIAM.
• El-Nabarawy I, Abdelbar AM and Wunsch DC (2013) Levenberg-Marquardt and Conjugate Gradient methods applied to a high-order neural network. In: Neural Networks (IJCNN), The 2013 International Joint Conference on, pp. 1–7.
• Grivell RM, Alfirevic Z, Gyte GM, et al. (2010) Antenatal cardiotocography for fetal assessment. The Cochrane database of systematic reviews, England (1): 1–48.
• Günther F and Fritsch S (2010) neuralnet: Training of neural networks. The R journal 2(1): 30–38.
• Hagan MT and Menhaj MB (1994) Training feedforward networks with the Marquardt algorithm. IEEE transactions on Neural Networks, IEEE 5(6): 989–993.
• Hagan MT, Demuth HB, Beale MH, et al. (2014) Neural network design. Martin Hagan. Available from: https://books.google.com.tr/books?id=4EW9oQEACAAJ.
• Hu YH and Hwang J-N (2001) Handbook of neural network signal processing. CRC press.
• Huang M-L and Yung-Yan H (2012) Fetal distress prediction using discriminant analysis, decision tree, and artificial neural network. Journal of Biomedical Science and Engineering 5: 526–533.
• Jezewski M, Czabanski R, Wrobel J, et al. (2010) Analysis of Extracted Cardiotocographic Signal Features to Improve Automated Prediction of Fetal Outcome. Biocybernetics and Biomedical Engineering 30(4): 29–47. Available from: http://ibib.waw.pl/images/ibib/grupy/Wydawnictwa-Tomy/dokumenty/2010/BBE_30_4_029_FT.pdf.
• Karabulut EM and Ibrikci T (2014) Analysis of Cardiotocogram Data for Fetal Distress Determination by Decision Tree Based Adaptive Boosting Approach. Computer and Communications 2(7): 32–37.
• Lichman M (2013) {UCI} Machine Learning Repository. Available from: http://archive.ics.uci.edu/ml.
• Luenberger DG, Ye Y and others (1984) Linear and nonlinear programming. Springer.
• Marquardt D (1963) An Algorithm for Least-Squares Estimation of Nonlinear Parameters. Journal of the Society for Industrial and Applied Mathematics, Society for Industrial and Applied Mathematics 11(2): 431–441. Available from: http://dx.doi.org/10.1137/0111030.
• Menai M, Mohder F and Al-mutairi F (2013) Influence of Feature Selection on Naïve Bayes Classifier for Recognizing Patterns in Cardiotocograms. Journal of Medical and Bioengineering 2(1): 66–70.
• Møller MF (1993) A scaled conjugate gradient algorithm for fast supervised learning. Neural networks, Elsevier 6(4): 525–533.
• Nunes I and Ayres-de-Campos D (2016) Computer analysis of foetal monitoring signals. Best Practice & Research Clinical Obstetrics & Gynaecology 30: 68–78. Available from: http://www.sciencedirect.com/science/article/pii/S1521693415000991.
• Ocak H (2013) A medical decision support system based on support vector machines and the genetic algorithm for the evaluation of fetal well-being. J Med Syst 37(2): 9913. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23321973.
• Onnia V, Tico M and Saarinen J (2001) Feature selection method using neural network. In: International Conference on Image Processing, pp. 513–516.
• Pinas A and Chandraharan E (2016) Continuous cardiotocography during labour: Analysis, classification and management. Best Practice & Research Clinical Obstetrics & Gynaecology 30: 33–47.
• Powell MJD (1977) Restart procedures for the conjugate gradient method. Mathematical programming, Springer 12(1): 241–254.
• Ravindran S, Jambek AB, Muthusamy H, et al. (2015) A novel clinical decision support system using improved adaptive genetic algorithm for the assessment of fetal well-being. Computational and mathematical methods in medicine, United States 2015: 283532.
• Riedmiller M and Braun H (1993) A direct adaptive method for faster backpropagation learning: The RPROP algorithm. In: Neural Networks, 1993., IEEE International Conference on, pp. 586–591.
• Sahin H and Subasi A (2015) Classification of the cardiotocogram data for anticipation of fetal risks using machine learning techniques. Applied Soft Computing 33: 231–238. Available from: http://www.sciencedirect.com/science/article/pii/S1568494615002653.
• Saini LM and Soni MK (2002) Artificial neural network-based peak load forecasting using conjugate gradient methods. IEEE Transactions on Power Systems, IEEE 17(3): 907–912.
• Sundar C, Chitradevi M and Geetharamani G (2012) Classification of cardiotocogram data using neural network based machine learning technique. International Journal of Computer Applications 47(14).
• Tomáš P, Krohová J, Dohnálek P, et al. (2013) Classification of cardiotocography records by random forest. In: 36th International Conference on Telecommunications and Signal Processing (TSP), pp. 620–923.
• Tongsong T, Iamthongin A, Wanapirak C, et al. (2005) Accuracy of fetal heart-rate variability interpretation by obstetricians using the criteria of the National Institute of Child Health and Human Development compared with computer-aided interpretation. Journal of Obstetrics and Gynaecology Research, Australia 31(1): 68–71.
• Wythoff BJ (1993) Backpropagation neural networks: a tutorial. Chemometrics and Intelligent Laboratory Systems, Elsevier 18(2): 115–155.
• Yılmaz E (2016) Fetal State Assessment from Cardiotocogram Data Using Artificial Neural Networks. Journal of Medical and Biological Engineering, Springer Berlin Heidelberg: 1–13. Available from: http://link.springer.com/10.1007/s40846-016-0191-3 (accessed 5 December 2016).