使用BNT学习动态贝叶斯网络的参数

我正尝试在Matlab中使用贝叶斯网络工具箱（BNT）创建动态贝叶斯网络。我一直在密切关注本教程，并用下面的代码结束：使用BNT学习动态贝叶斯网络的参数

T=2; 
names = {'X1', 'X2', 'X3', 'X4', 'X5', 'X6', 'X7', 'X8', 'X9', 'X10', 'X11', 'X12'}; 
ss = length(names); 
onodes=[1 2 3 4 5 6 7 8 9 10 11 12]; 
intrac = { 
    'X1', 'X2'; 
    'X3', 'X4'; 
    'X6', 'X7'; 
    'X7', 'X8'; 
    'X10', 'X11'; 
    'X10', 'X12'; 
    'X11', 'X12'; 
    }; 
[intra, names] = mk_adj_mat(intrac, names, 1); 
interc = { 
    'X1', 'X1'; 
    'X2', 'X2'; 
    'X3', 'X3'; 
    'X4', 'X4'; 
    'X5', 'X5'; 
    'X6', 'X6'; 
    'X7', 'X7'; 
    'X8', 'X8'; 
    'X9', 'X9'; 
    'X10', 'X10'; 
    'X11', 'X11'; 
    'X12', 'X12'; 
    }; 
inter = mk_adj_mat(interc, names, 0); 
dnodes = 1:ss; 
ns = [3 3 3 2 4 5 2 2 4 5 4 5]; 
ns(stringmatch('X1',names))=3; 
ns(stringmatch('X2',names))=3; 
ns(stringmatch('X3',names))=3; 
ns(stringmatch('X4',names))=2; 
ns(stringmatch('X5',names))=4; 
ns(stringmatch('X6',names))=5; 
ns(stringmatch('X7',names))=2; 
ns(stringmatch('X8',names))=2; 
ns(stringmatch('X9',names))=4; 
ns(stringmatch('X10',names))=5; 
ns(stringmatch('X11',names))=4; 
ns(stringmatch('X12',names))=5; 

bnet = mk_dbn(intra, inter, ns, 'discrete', dnodes); 
for i=1:2*ss 
    bnet.CPD{i} = tabular_CPD(bnet, i); 
end 
allcases=importdata('Myfile.txt',','); 
ncases = size(allcases.data,1); 
cases = cell(1, ncases); 
for i=1:ncases 
    cases{i} = cell(ss,T); 
    dataentry=allcases.data(i,:); 
    ev=uint8([dataentry(1:12)' dataentry(13:24)']); 
    cases{i} = num2cell(ev); 
end 
engine2 = smoother_engine(jtree_2TBN_inf_engine(bnet)); 
%engine2 = jtree_2TBN_inf_engine(bnet); 
[bnt, LL, engine2] = learn_params_dbn_em(engine2, cases, 'max_iter', 20);

当我尝试运行它，以下错误显示出来：

Subscript indices must either be real positive integers or logicals. 

Error in discrete_CPD/convert_to_table (line 14) 
T = CPT(index{:}); 

Error in discrete_CPD/convert_to_pot (line 20) 
    T = convert_to_table(CPD, domain, evidence); 


Error in jtree_2TBN_inf_engine/fwd1 (line 11) 
    CPDpot{n} = convert_to_pot(bnet.CPD{e}, engine.pot_type, fam(:), ev); 

Error in smoother_engine/enter_evidence (line 12) 
[f{1}, ll(1)] = fwd1(engine.tbn_engine, ev(:,1), 1); 

Error in learn_params_dbn_em>EM_step (line 131) 
    [engine, ll] = enter_evidence(engine, evidence); 

Error in learn_params_dbn_em (line 82) 
    [engine, loglik, logpost] = EM_step(engine, evidence, temperature);

有人可以帮我这个错误？我认为这与我创建表格CPD的方式有关，因为当我查看bnet中的CPD时，他们都是1 * 1表格。我的CPD的创建方式与教程相同。提前谢谢！

来源

2015-04-02 Junfei Wang

更新：我刚才发现这个错误与表格CPD无关。 – 2015-04-02 19:48:25

更新＃2：我对整个网络进行了采样，它使用采样数据很好地学习了参数。所以我认为我的问题在我的数据集中。现在看看它 – 2015-04-02 21:17:10