Mixture Density Network Estimation of Continuous Variable Maximum Likelihood Using Discrete Training Samples
share
Mixture Density Networks (MDNs) can be used to generate probability density functions of model parameters θ given a set of observables 𝐱. In some applications, training data are available only for discrete values of a continuous parameter θ. In such situations a number of performance-limiting issues arise which can result in biased estimates. We demonstrate the usage of MDNs for parameter estimation, discuss the origins of the biases, and propose a corrective method for each issue.
READ FULL TEXT
