BNW using Octave instead of Matlab.

This version of BNW should perform the same as the original version. The only difference is that it uses Octave instead of Matlab when running BayesNet Toolbox during parameter learning.

I am calling this BNW_1.02. It can be accessed at:
compbio.uthsc.edu/BNW_1.02

1 files changed, 30 insertions, 0 deletions

diff --git a/sourcecodes/bnt-master/BNT/potentials/Tables/mult_by_table2.m b/sourcecodes/bnt-master/BNT/potentials/Tables/mult_by_table2.m
new file mode 100644
index 00000000..92fb4ac7
--- /dev/null
+++ b/sourcecodes/bnt-master/BNT/potentials/Tables/mult_by_table2.m
@@ -0,0 +1,30 @@
+function bigT = mult_by_table2(bigT, bigdom, bigsz, smallT, smalldom, smallsz)
+% MULT_BY_TABLE 
+% bigT = mult_by_table(bigT, bigdom, bigsz, smallT, smalldom, smallsz)
+%
+
+%Ts = extend_domain_table(smallT, smalldom, smallsz, bigdom, bigsz);
+%bigT(:) = bigT(:) .* Ts(:); % must have bigT(:) on LHS to preserve shape
+
+% extend_domain_table has a lot of overhead for small tables,
+% since it calls myreshape and myrepmat, which check for 1 dimensional case.
+% Here, we check up front.
+
+if length(bigdom)==1 % vector
+  bigT = bigT .* smallT; % smallT can be scalar or vector
+else
+  if (length(bigsz) == length(smallsz)) & all(bigsz == smallsz)
+    bigT = bigT .* smallT;
+  else
+    map = find_equiv_posns(smalldom, bigdom);
+    sz = ones(1, length(bigdom));
+    sz(map) = smallsz;
+    smallT = reshape(smallT, sz); % add dimensions of size 1 for missing domain
+    % we can use reshape instead of myreshape, because we know length(sz)>1
+    sz = bigsz;
+    sz(map) = 1; % don't replicate along small domain, which is shared
+    % we can use repmat instead of myrepmat, because we know length(sz)>1
+    smallT = repmat(smallT, sz(:)');
+    bigT(:) = bigT(:) .* smallT(:);
+  end
+end