Hybrid Optimization Algorithm for Gender Prediction from Anthropometric Attributes Using Machine Learning

Gender determination holds a great significance in medical and legal analysis. The designated authorities make a judgment based on the analysis of some of the anatomical structures of human body. Anthropologists use skeletal biomarkers that are different depending on the gender to find the sex. It is nearly every bone of the skeleton that has been used in a great deal of research to see the accuracy of sex estimation through the gender determination method, and the accuracy of gender determination has been evaluated through a number of comparison studies conducted across various populations. Pre-processing, deriving features, and classifying are the key steps in the process of predicting gender from anthropometric features. This work presents a hybrid optimization algorithm that uses two main techniques which are PSO (particle swarm optimization) and genetic algorithms. For classifying, a random forest model is utilized. The evaluated model is implemented, and its performance has been evaluated against other standard approaches (e.g. Naïve bayes, KNN, Random Forest, Decision Tree).