Classification of Obese and Healthy Children Based on Machine Learning Algorithms using Choroidal Thickness Features

Abstract Purpose: To analyse the effect of macular choroidal thickness (MCT) and peripapillary choroidal thickness (PPCT) on the classification of obese and healthy children by comparing the performance of Random Forest (RF) or Support Vector Machine (SVM), and Multilayer Perceptrons (MLP) algorithms. Methods: 59 obese children and 35 healthy children aged 6 to 15 years were studied in this prospective comperative study using optical coherence tomography. MCT and PPCT were measured at 500 μm, 1000 μm, and 1500 μm distances from fovea and optic disc. Three different feature selection algorithms were used to determine the most prominent features of all extracted features. The classification efficiency of the extracted features was analyzed using RF, SVM, and MLP algorithms, demonstrating their efficacy for distinguishing obese from healthy children. The precision and reliability of measurements were assessed using Kappa analysis.Result: Correlation Feature Selection algorithm produced the most successful classification results among the different feature selection methods. The most prominent features for distinguishing the obese and healthy groups from each other were PPCT temporal 500 μm, PPCT temporal 1500 μm, PPCT nasal 1500 μm, PPCT inferior 1500 μm, and subfoveal MCT. The classification rates for RF, SVM, and MLP algorithms were 98.6%, 96.8% and 89%, respectively. Conclusion: Obesity-related metabolic alterations have an effect on the choriocapillaries of children, particularly in the subfoveal region and the outer semi-circle at 1500 μm from optic disc head. Both the RF and SVM algorithms are effective and accurate at classifying obese and healthy children.