Gender Classification of Height-Weight 2D Data Using Bayesian Method with Supervised Parameter Estimation of Normal Distribution

In Professor Zhang Xuegong's first pattern recognition assignment, we implemented gender classification on two-dimensional height-weight data using Bayesian method with supervised parameter estimation based on normal distribution. Specifically, we first collected a dataset containing height and weight measurements, then performed classification using the Bayesian approach that relies on supervised parameter estimation for normal distributions. The implementation involves calculating maximum likelihood estimates (MLE) for mean vectors and covariance matrices for each gender class, followed by computing posterior probabilities using Bayes' theorem. Through analyzing the height-weight data patterns, we can determine an individual's gender based on the Bayesian classification results. This method plays a significant role in pattern recognition field, helping us better understand and classify data by implementing probabilistic decision boundaries and density estimation techniques. The key algorithmic components include multivariate normal density calculation, prior probability estimation, and discriminant function implementation for optimal classification performance.