Abstract
We present an algorithm for estimating parameters in a mixture-of-regressions model in which the errors are assumed to be independent and identically distributed but no other assumption is made. This model is introduced as one of several recent generalizations of the standard fully parametric mixture of linear regressions in the literature. A sufficient condition for the identifiability of the parameters is stated and proved. Several different versions of the algorithm, including one that has a provable ascent property, are introduced. Numerical tests indicate the effectiveness of some of these algorithms.
| Original language | English |
|---|---|
| Pages (from-to) | 19-38 |
| Number of pages | 20 |
| Journal | Journal of Nonparametric Statistics |
| Volume | 24 |
| Issue number | 1 |
| DOIs | |
| State | Published - Mar 2012 |
Bibliographical note
Funding Information:This research was supported by NSF Award SES-0518772. We thank the reviewers for numerous helpful comments.
Keywords
- EM algorithm
- density estimation
- finite mixture model
- identifiability
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty