Physics and Astronomy
5 years ago

New method simplifies learning graphical models for better data analysis.

1 view
arXiv (Cornell University)
Composite Likelihood Estimation for Restricted Boltzmann machines
Muneki Yasuda, S. Kataoka, Yuji Waizumi, Kazuyuki Tanaka

Paper Summary

Learning the parameters of graphical models is difficult, so scientists use a method called composite likelihood estimation to make it easier. This method is a statistical approximation that helps estimate the parameters more accurately. In this study, researchers developed a new composite likelihood method and tested it on restricted Boltzmann machines. The results showed that the new method is effective in estimating the parameters of these models.

Related papers

Theme background image
Engineering
8 years ago

New method improves accuracy of time series predictions for better decision-making.

Process Fault Detection
Theme background image
Mathematics
8 years ago

Large inspection size inefficient for low failure rate in exponential distribution.

Skew Distributions
Statistical Control
Missing Data
Theme background image
Mathematics
3 years ago

New method for accurate probability density estimation without unstable estimates!

Regularization Methods
Multicollinearity
System Identification