|
Activity Number:
|
340
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Quality and Productivity Section
|
|
Abstract #319771
|
|
|
Title:
|
Mixture of Forecasting Models with an Application on Solar Energy Forecasting
|
|
Author(s):
|
YoungDeok Hwang* and Siyuan Lu and Eric Wang
|
|
Companies:
|
IBM T. J. Watson Research Center and IBM T. J. Watson Research Center and Duke University
|
|
Keywords:
|
mixture model ;
computer models
|
|
Abstract:
|
This paper introduces a mixture modeling approach appropriate for a large monitoring network in which measurements and associated covariates are taken over time. The proposed approach is developed for unveiling the clustering structure of data, which can help the prediction accuracy as well as understanding of physical system. The modeling framework can incorporate a complex structure of the data with an efficient algorithm for real-time model fitting. We also provide a criterion for selection of the optimal number of clusters. Our approach is illustrated using simulation studies and a solar energy forecasting problem.
|
Authors who are presenting talks have a * after their name.
