All Times ET
Keywords: EM-algorithm, Laplace distribution, Maximum likelihood estimation, Mixture models, Renewable energy
We introduce a new stochastic model connected with mixtures of uniform and Laplace distributions. We present basic properties of this model and discuss its potential applications for classification and outlier detection. We also discuss computational approaches to parameter estimation through expectation-maximization schemes and present a data example involving modeling renewable energy (solar and wind) output forecast errors.