Unsupervised weight parameter estimation for exponential mixture distribution based on symmetric Kullback-Leibler divergence

Masato Uchida

Abstract

When there are multiple component predictors, it is promising to integrate them into one predictor for advanced reasoning. If each component predictor is given as a stochastic model in the form of probability distribution, an exponential mixture of the component probability distributions provides a good way to integrate them. However, weight parameters used in the exponential mixture model are difficult to estimate if there is no data for performance evaluation. As a suboptimal way to solve this problem, weight parameters may be estimated so that the exponential mixture model should be a balance point that is defined as an equilibrium point with respect to the distance from/to all component probability distributions. In this paper, we propose a weight parameter estimation method that represents this concept using a symmetric Kullback-Leibler divergence and discuss the features of this method.
BibTeX
@article{uchida2015unsupervised,
  author = {Masato Uchida},
  title = {{Unsupervised weight parameter estimation for exponential mixture distribution based on symmetric Kullback-Leibler divergence}},
  journal = {IEICE Trans. Fundam. Electron. Commun. Comput. Sci.},
  year = {2015},
  url = {https://openreview.net/forum?id=MsOFcJFvU3},
}

Loading related content...

Figure