A Novel Sum-Product Algorithm on Factor Graphs
Date Issued
2004
Date
2004
Author(s)
Pan, Ting-Jian
DOI
en-US
Abstract
Factor graphs provide a unified and efficient graphical perspective for various well-known graphical models as we see in communication, signal processing, and artificial intelligence communities, such as LDPC codes, hidden Markov models, Bayesian networks, Markov random fields, and certain Fourier transforms. Due to the wide applications of factor graphs, the design of the coding algorithm on a factor graph is very important. The main aim of this paper is to review the current coding algorithm on a factor graph, the sum-product algorithm, and to provide an efficient, fast, and flexible sum-product algorithm for solving both general situations and some constrained situations of marginalization problems.
Subjects
隱藏式馬可夫模型
和積演算法
傅立葉轉換
低密度同位元檢查碼
馬可夫隨機場模型
貝氏網路
因子圖
Fourier Transform
Factor Graph
Hidden Markov Model
Low-Density Parity-Check Code
Bayesian Network
Sum-Product Algorithm
Markov Random Field
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-93-R91922055-1.pdf
Size
23.31 KB
Format
Adobe PDF
Checksum
(MD5):111bc8556ad89efbcec05794cadf4b25