Mercurial > hg > camir-aes2014
comparison toolboxes/FullBNT-1.0.7/bnt/examples/dynamic/HHMM/Motif/fixed_args_mk_motif_hhmm.m @ 0:e9a9cd732c1e tip
first hg version after svn
author | wolffd |
---|---|
date | Tue, 10 Feb 2015 15:05:51 +0000 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
-1:000000000000 | 0:e9a9cd732c1e |
---|---|
1 function bnet = fixed_args_mk_motif_hhmm(motif_length, motif_pattern, background_char) | |
2 % | |
3 % BNET = MK_MOTIF_HHMM(MOTIF_LENGTH) | |
4 % Make the following HHMM | |
5 % | |
6 % S2 <----------------------> S1 | |
7 % | | | |
8 % | | | |
9 % M1 -> M2 -> M3 -> end B1 -> end | |
10 % | |
11 % where Mi represents the i'th letter in the motif | |
12 % and B is the background state. | |
13 % Si chooses between running the motif or the background. | |
14 % The Si and B states have self loops (not shown). | |
15 % | |
16 % The transition params are defined to respect the above topology. | |
17 % The background is uniform; each motif state has a random obs. distribution. | |
18 % | |
19 % BNET = MK_MOTIF_HHMM(MOTIF_LENGTH, MOTIF_PATTERN) | |
20 % In this case, we make the motif submodel deterministically | |
21 % emit the motif pattern. | |
22 % | |
23 % BNET = MK_MOTIF_HHMM(MOTIF_LENGTH, MOTIF_PATTERN, BACKGROUND_CHAR) | |
24 % In this case, we make the background submodel | |
25 % deterministically emit the specified character (to make the pattern | |
26 % easier to see). | |
27 | |
28 if nargin < 2, motif_pattern = []; end | |
29 if nargin < 3, background_char = []; end | |
30 | |
31 chars = ['a', 'c', 'g', 't']; | |
32 Osize = length(chars); | |
33 | |
34 motif_length = length(motif_pattern); | |
35 Qsize = [2 motif_length]; | |
36 Qnodes = 1:2; | |
37 D = 2; | |
38 transprob = cell(1,D); | |
39 termprob = cell(1,D); | |
40 startprob = cell(1,D); | |
41 | |
42 % startprob{d}(k,j), startprob{1}(1,j) | |
43 % transprob{d}(i,k,j), transprob{1}(i,j) | |
44 % termprob{d}(k,j) | |
45 | |
46 | |
47 % LEVEL 1 | |
48 | |
49 startprob{1} = zeros(1, 2); | |
50 startprob{1} = [1 0]; % always start in the background model | |
51 | |
52 % When in the background state, we stay there with high prob | |
53 % When in the motif state, we immediately return to the background state. | |
54 transprob{1} = [0.8 0.2; | |
55 1.0 0.0]; | |
56 | |
57 | |
58 % LEVEL 2 | |
59 startprob{2} = 'leftstart'; % both submodels start in substate 1 | |
60 transprob{2} = zeros(motif_length, 2, motif_length); | |
61 termprob{2} = zeros(2, motif_length); | |
62 | |
63 % In the background model, we only use state 1. | |
64 transprob{2}(1,1,1) = 1; % self loop | |
65 termprob{2}(1,1) = 0.2; % prob transition to end state | |
66 | |
67 % Motif model | |
68 transprob{2}(:,2,:) = mk_leftright_transmat(motif_length, 0); % no self loops | |
69 termprob{2}(2,end) = 1.0; % last state immediately terminates | |
70 | |
71 | |
72 % OBS LEVEl | |
73 | |
74 obsprob = zeros([Qsize Osize]); | |
75 if isempty(background_char) | |
76 % uniform background model | |
77 obsprob(1,1,:) = normalise(ones(Osize,1)); | |
78 else | |
79 % deterministic background model (easy to see!) | |
80 m = find(chars==background_char); | |
81 obsprob(1,1,m) = 1.0; | |
82 end | |
83 | |
84 if gen_motif | |
85 % initialise with true motif (cheating) | |
86 for i=1:motif_length | |
87 m = find(chars == motif_pattern(i)); | |
88 obsprob(2,i,m) = 1.0; | |
89 end | |
90 else | |
91 obsprob(2,:,:) = mk_stochastic(ones(motif_length, Osize)); | |
92 end | |
93 | |
94 Oargs = {'CPT', obsprob}; | |
95 | |
96 [bnet, Qnodes, Fnodes, Onode] = mk_hhmm('Qsizes', Qsize, 'Osize', Osize, 'discrete_obs', 1, ... | |
97 'Oargs', Oargs, 'Ops', Qnodes(1:2), ... | |
98 'startprob', startprob, 'transprob', transprob, 'termprob', termprob); | |
99 |