Parameter estimation of Gaussian mixture model utilizing boundary data

Gaussian mixture model is a statistical model that represents a distribution of data correctly, which can be used for prediction, monitoring, segmentation, discrimination, clustering, recognition, etc. The parameters of the Gaussian mixture model are usually estimated from given sample data by the Expectation Maximization algorithm. However, when the number of attributes of the data is large, the parameters cannot be estimated correctly. In this paper, we propose a novel approach for estimating the parameters of the Gaussian mixture model by utilizing the sample data located on the boundary of regions defined by the component density functions. The effectiveness of the proposed method is confirmed by some experiments.