The potential of SMOTHE to enhance the precision of gender prediction: an investigation of artificial neural networks with cephalometry determination

Abstract

Background: When creating models utilizing artificial neural networks (ANN), it is crucial to consider the quantity of training data and the distribution of data, particularly when making gender predictions. Objectives: This study seeks to determine the potential impact of using Synthetic Minority Oversampling Technique (SMOTE) on gender prediction using ANN model. Material and Method: The current study utilized a dataset consisting of 297 Indonesian cephalometric measurements, comprising 229 samples from females and 68 samples from males. Web ceph is using for measures parameters SNA angle, mandibular length, mandibular angle, and SGA angle and diagnosis. Data processing and model ANN creation were carried out using Python. Result: The gender identification accuracy of the artificial neural network (ANN) model is 87% for females and 0% for males, resulting in an overall average accuracy of 78%. When using SMOTE, the accuracy is 22%, with 0% accuracy for females and 37% accuracy for males. However, when using SMOTE and normalization, the accuracy increases to 71%, with 82% accuracy for females and 30% accuracy for males. The accuracy of normalization without SMOTE is 76%, with 86% accuracy for females and 14% accuracy for males. Conclusion: Research has proven the efficacy of SMOTE in improving the classification of malematrices. Nevertheless, the study reveals that the overall accuracy results of SMOTE are suboptimal in comparison to the absence of SMOTE and normalization. The application of data balancing strategies is necessary in order to achieve optimal accuracy in gender prediction when ANN, and other parameters must be applied.