Anda belum login :: 23 Nov 2024 05:23 WIB
Home
|
Logon
Hidden
»
Administration
»
Collection Detail
Detail
A Latent Variable Model Learning Method using Bayesian Networks
Oleh:
Lee, Su-Dong
;
Jun, Chi-Hyuck
Jenis:
Article from Proceeding
Dalam koleksi:
The 14th Asia Pacific Industrial Engineering and Management Systems Conference (APIEMS), 3-6 December 2013 Cebu, Philippines
,
page 1-11.
Topik:
Latent variable model
;
Bayesian networks
;
Variable clustering
;
Causal discovery
;
Partial least squares
Fulltext:
1279.pdf
(716.36KB)
Isi artikel
It is sometimes advantageous to analyze the underlying relations among variablesin terms of the latent variables not of the original variables.To discover and construct the latent variable model, we propose a procedure including the following steps: 1) to extract latent variables, 2) to construct the causal structure of latent variables, and 3) estimate the parameters of the structure. For latent variable extraction, variable clustering using principal component analysis is conducted to group the related manifest variables which share a common latent variable. Then, the causal structure among the latent variables is constructed by using Bayesian networks. Finally, the parameters of the learned structure are estimated via a partial least squares algorithm. Through an experiment with synthetic data, the proposed method is validated.
Opini Anda
Klik untuk menuliskan opini Anda tentang koleksi ini!
Kembali
Process time: 0.015625 second(s)