Debian Codegraph @

texmf-dist/doc/latex/tikz-bayesnet/example.tex

raw
1 % example.tex
2 %
3 % Copyright (C) 2010,2011 Laura Dietz
4 % Copyright (C) 2012 Jaakko Luttinen
5 %
6 % This file may be distributed and/or modified
7 %
8 % 1. under the LaTeX Project Public License and/or
9 % 2. under the GNU General Public License.
10 %
11 % See the files LICENSE_LPPL and LICENSE_GPL for more details.
12
13 \documentclass[a4paper]{article}
14
15 \usepackage{tikz}
16 \usetikzlibrary{bayesnet}
17 %\pgfrealjobname{example} % name of this file
18
19 \title{Graphical Models in Tikz}
20 \author{Laura Dietz, Jaakko Luttinen}
21
22 \begin{document}
23
24 \maketitle
25
26 TikZ examples for graphical models (Bayesian networks) and directed
27 factor graphs \cite{Dietz:2010}.
28
29 % A table of node types
30 \begin{table}[ht]
31 \caption{Node types}
32 \begin{center}
33 \begin{tabular}{llc}
34 Type & Syntax & Output
35 \\
36 \hline
37 Latent variable &
38 \texttt{\textbackslash node[latent]} &
39 \tikz{ %
40 \node[latent] {$x$}; %
41 }
42 \\
43 Observed variable &
44 \texttt{\textbackslash node[obs]} &
45 \tikz{ %
46 \node[obs] {$y$}; %
47 }
48 \\
49 Deterministic &
50 \texttt{\textbackslash node[det]} &
51 \tikz{ %
52 \node[det] {dot} ; %
53 }
54 \\
55 Constant &
56 \texttt{\textbackslash node[const]} &
57 \tikz{ %
58 \node[const] {$a$}; %
59 }
60 \\
61 Factor &
62 \texttt{\textbackslash node[factor]} &
63 \tikz{ %
64 \node[factor] [label=$\mathcal{N}$] {}; %
65 }
66 \\
67 Factor with nodes &
68 &
69 \tikz{ %
70 \node[obs] (y) {$y$} ; %
71 \node[latent, left=of y, yshift=0.5cm] (mu) {$\mu$} ; %
72 \node[latent, left=of y, yshift=-0.5cm] (tau) {$\tau$} ; %
73 \factor[left=of y] {y-factor} {$\mathcal{N}$} {} {};
74 \factoredge {mu,tau} {y-factor} {y} ; %
75 }
76 \\
77 Plate &
78 \texttt{\textbackslash plate} &
79 \tikz{ %
80 \node[latent] (x) {$x_m$}; %
81 \plate {} {(x)} {$m \in \mathcal{M}$}; %
82 }
83 \\
84 Gate &
85 &
86 \tikz{
87 % Nodes
88 \node[obs] (k) {$k$}; %
89 \node[latent, above=2 of k] (l) {$\lambda$}; %
90 \factor[above=0.8 of k] {k-f} {Multi} {} {}; %
91 \node[latent, right=of k-f] (paa) {$\phi$}; %
92 %\node[latent, right=of k-f] (p) {$\phi$}; %
93 % Connections
94 \factoredge {paa} {k-f} {k} ; %
95 % Gate
96 \gate {} {(k-f)(k-f-caption)} {l} ; %
97 }
98 \end{tabular}
99 \end{center}
100 \end{table}
101
102
103 % Simple Bayesian network
104 \begin{figure}[ht]
105 \begin{center}
106 \begin{tabular}{cc}
107 \input{model_pca} &
108 \input{model_pca2}
109 \end{tabular}
110 \end{center}
111 \caption{PCA model as a Bayesian network and a directed factor
112 graph.}
113 \end{figure}
114
115 % Latent Dirichlet allocation
116 \begin{figure}[ht]
117 \begin{center}
118 \input{model_lda}
119 \end{center}
120 \caption{Latent Dirichlet allocation as directed factor graph.}
121 \end{figure}
122
123 % Citation influence model
124 \begin{figure}[ht]
125 \begin{center}
126 \input{model_citation_influence}
127 \end{center}
128 \caption{Citation influence model with own topics \cite{Dietz:2007}
129 as directed factor graph.}
130 \end{figure}
131
132 \clearpage
133
134 \begin{thebibliography}{9}
135
136 \bibitem{Dietz:2010}
137 Laura Dietz,
138 \emph{Directed Factor Graph Notation for Generative Models}.
139 Technical Report. 2010
140
141 % Laura Dietz, Steffen Bickel, Tobias Scheffer.
142 % Unsupervised Prediction of Citation Influences.
143 % In: Proceedings of International Conference on Machine Learning. 2007
144 \bibitem{Dietz:2007}
145 Laura Dietz, Steffen Bickel, Tobias Scheffer,
146 \emph{Unsupervised Prediction of Citation Influences}.
147 In: Proceedings of International Conference on Machine
148 Learning. 2007
149
150
151 \end{thebibliography}
152
153 \end{document}
154
155 %%% Local Variables:
156 %%% mode: tex-pdf
157 %%% TeX-master: t
158 %%% End: