Artificial Neural Network for Multiclass Recognition and its Application to the Thyroid Functional State
Keywords:Artificial neural network, Classification, Ultrasound image processing, Regularization, Inverse problem
Background. Development of automated diagnostic requires selection and improvement of appropriate machine learning methods, in particular multiclass recognition. Artificial Neural Networks (ANN) of various architecture are considered as an approach to the problem.
Objective. The goal is to analyze and compare performance of ANN-based classifiers on various datasets for further improvement of model selection strategy.
Methods. ANN-based models of the distribution of class labels in terms of predictor features are constructed, trained and validated for datasets of clinical records. Varying training algorithms for multi-layer perceptrons, Kohonen neural network, linear functional strategy with multi-parameters regularization are considered.
Results. Performance of the classifiers is compared in terms of accuracy, sensitivity, and specificity. Linear functional strategy classifier outperforms the other with more complex ANN-architecture and exhibits relative steadiness to overfitting. Performance of Kohonen neural network on large dataset exceeds 90 % in terms of specificity for each class, withal sensitivity for distinct classes is more than 95 %.Conclusions. The understanding of the strengths and limitations of each method is crucial for careful choice of ANN-based classifier, particularly its architecture, regularization and training algorithm.
J.A. Noble, “Ultrasound image segmentation and tissue characterization”, Proc. In. Mech. Eng. H, J. Eng. Medicine, vol. 223, pp. 1–10, 2009. doi: 10.1243/09544119JEIM604
E.C. Kyriacou et al., “A review of noninvasive ultrasound image processing methods in the analysis of carotid plaque morphology for the assessment of stroke risk”, IEEE Trans. Inform. Technol. Biomed., vol. 14, iss. 4, pp. 1027–1038, 2010. doi: 10.1109/TITB.2010.2047649
G. Slabaugh et al., “Statistical region-based segmentation of ultrasound images”, Ultrasound Med. Biol. vol. 35, no. 5, pp. 781–795, 2009. doi: 10.1016/j.ultrasmedbio.2008.10.014
M.K. Novoselets et al., “Ultrasound introscopic image quantitative characteristics for medical diagnostics and refinements of physical noise rise reasons”, Proc. SPIE, Medical Imaging: Physics of Medical Imaging, vol. 2163, 1994, pp. 442–446. doi: 10.1117/12.174281
R.V. Byvalkevich et al., “Tools for ultrasonic diagnostic image classification”, in Proc. 8th IEEE Int. Conf. Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, Warsaw, Poland, 2015, pp. 977–981. doi: 10.1109/ IDAACS.2015.7341450
G. Kriukova et al., “A linear functional strategy for regularized ranking”, Neural Networks, vol. 73, pp. 26–35, 2016. doi: 10.1016/j.neunet.2015.08.012
P. Tkachenko et al., “Prediction of nocturnal hypoglycemia by an aggregation of previously known prediction approaches: proof of concept for clinical application”, Comput. Methods Programs Biomed, vol. 134, pp. 179–186, 2016. doi: 10.1016/j.cmpb.2016.07.003
R.M. Haralick et al., “Textural features for image classification”, IEEE Trans. Syst. Man Cybern., vol. SMC-3, no. 6, pp. 610–621, 1973. doi: 10.1109/TSMC.1973.4309314
S. Aksenov and V. Novoseltsev, Neural Networks Organization and Using (Methods and Techniques). Tomsk, Russia, 2006 (in Russian).
M. Lichman, UCI Machine Learning Repository. Irvine, CA: University of California, School of Information and Computer Science, 2013. Available: http://archive.ics.uci.edu/ml/datasets/Thyroid+Disease
W. Schiffmann et al., “Comparison of optimized backpropagation algorithms”, in Proc. Europ. Symp. Artificial Neural Networks (ESANN'93), M. Verleysen Ed., Brussels, Belgium: de Facto Press, 1993, pp. 97–104.
E. Alba and F. Chicano, Generic Algorithms, in Metaheuristic Procedures for Training Neural Networks (Operations Research/ Computer Science Interfaces Series 35), E. Alba and R. Marti, Eds. Springer Science & Business Media, 2006, pp. 120–138.
I. Osorio et al., “Performance reassessment of a real-time seizure-detection algorithm on long ECoG series”, Epilepsia, vol. 43, no. 12, 2002, pp. 1522–1535. doi: 10.1046/j.1528-1157.2002.11102.x
W. Christian, Automatic Seizure-Detection in Rats using ECoG: A Comparison of Methods. Lübeck, Germany, 2013.
A. Salnikov et al., “Integrated grid environment for massive distributed computing in neuroscience”, in Proc. 6th IEEE Int. Conf. Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications, Prague, Czech Republic, 2011, pp. 198–202. doi: 10.1109/IDAACS.2011.6072740
V.O. Gaidar and O.O. Sudakov, “Archiving and analysis of electroencephalograms in Ukrainian grid: the first application”, in Proc. 8th IEEE Int. Conf. IDAACS, Warsaw, Poland, 2015, pp. 961–965. doi: 10.1109/IDAACS.2015.7341447
LicenseCopyright (c) 2017 NTUU KPI Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under CC BY 4.0 that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work