apaam/mlpack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

updated code so that user can specify number of ICs to find via num_of_IC. Also, fixed bug in very final step where we forgot to transpose unmixing matrix W before multiplying it to data

Browse Source
This commit is contained in:
tekhnofiend
2010-05-24 21:53:37 +00:00
parent 9713eafc86
commit 4f475b8855
2 changed files with 11 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastlib/trunk/mlpack/fastica/fastica.h
+10 -2
View File
@@ -33,6 +33,8 @@ const fx_entry_doc fastica_entries[] = {
"Independent component recovery approach: 'deflation' or 'symmetric'.\n"},
{"nonlinearity", FX_PARAM, FX_STR, NULL,
"Nonlinear function to use: 'logcosh', 'gauss', 'kurtosis', or 'skew'.\n"},
{"num_of_IC", FX_PARAM, FX_INT, NULL,
" Number of independent components to find: integer between 1 and dimensionality of data.\n"},
{"fine_tune", FX_PARAM, FX_BOOL, NULL,
"Enable fine tuning.\n"},
{"a1", FX_PARAM, FX_DOUBLE, NULL,
@@ -643,7 +645,13 @@ class FastICA {
//const index_t first_eig_ = fx_param_int(module_, "first_eig", 1);
// for now, the last eig must be d, and num_of IC must be d, until I have time to incorporate PCA into this code
//const index_t last_eig_ = fx_param_int(module_, "last_eig", d);
num_of_IC_ = d; //fx_param_int(module_, "num_of_IC", d);
num_of_IC_ = fx_param_int(module_, "num_of_IC", d);
if(num_of_IC_ < 1 || num_of_IC_ > d) {
printf("ERROR: num_of_IC = %d must be >= 1 and <= dimensionality of data",
num_of_IC_);
return SUCCESS_FAIL;
}
fine_tune_ = fx_param_bool(module_, "fine_tune", false);
a1_ = fx_param_double(module_, "a1", 1);
a2_ = fx_param_double(module_, "a2", 1);
@@ -1366,7 +1374,7 @@ class FastICA {
FixedPointICA(X_whitened, whitening_matrix, W);
if(ret_val == SUCCESS_PASS) {
la::MulInit(*W, X(), Y);
la::MulTransAInit(*W, X(), Y);
}
else {
Y -> Init(0,0);
fastlib/trunk/mlpack/fastica/fastica_main.cc
+1
View File
@@ -24,6 +24,7 @@
* @param seed = (long) seed to the random number generator (clock() + time(0))
* @param approach = {deflation, symmetric} (deflation)
* @param nonlinearity = {logcosh, gauss, kurtosis, skew} (logcosh)
* @param num_of_IC = integer constant for number of independent components to find (dimensionality of data)
* @param fine_tune = {true, false} (false)
* @param a1 = numeric constant for logcosh nonlinearity (1)
* @param a2 = numeric constant for gauss nonlinearity (1)