Learning mixture models with the regularized latent maximum entropy principle

• Published in: IEEE transactions on neural networks Vol. 15; no. 4; pp. 903 - 916
• Main Authors: Shaojun Wang, Schuurmans, D., Fuchun Peng, Yunxin Zhao
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.07.2004
• Subjects: Algorithms; Artificial Intelligence; Computer science; Computer Simulation; Decision Support Techniques; Entropy; Estimating; Inference; Inference algorithms; Information Storage and Retrieval - methods; Information Theory; Iterative algorithms; Learning; Machine learning; Maximization; Maximum entropy; Maximum likelihood estimation; Models, Statistical; Neural networks; Neural Networks (Computer); Parametric statistics; Pattern Recognition, Automated; Probability Learning; Robustness; State estimation; Yield estimation

This paper presents a new approach to estimating mixture models based on a recent inference principle we have proposed: the latent maximum entropy principle (LME). LME is different from Jaynes' maximum entropy principle, standard maximum likelihood, and maximum a posteriori probability estimati...