confidence

fluents/lib/blmfuncs.py

@@ -390,7 +390,7 @@ class LPLS(Model):
         """
         aopt = self.model['aopt']
         if opt['calc_conf']:
-            Wx, Wz = lpls_jk(self.model['X'], self.model['Y'], self.model['Z'], aopt, n_sets)
+            Wx, Wz = lpls_jk(self._data['X'], self._data['Y'], self._data['Z'], aopt, opt['n_sets'], opt['xz_alpha'])
             Wcal = self.model['W'][:,:aopt]
             Lcal = self.model['L'][:,:aopt]
             # ensure that Wcal is scaled
@@ -669,8 +669,8 @@ class LplsOptions(Options):
         opt['center'] = True
         opt['center_mth'] = [2, 0, 1]
         opt['scale'] = 'scores'
-        opt['calc_conf'] = False
+        opt['calc_conf'] = True
-        opt['n_sets'] = 7
+        opt['n_sets'] = 20
         opt['strict'] = False
         opt['p_center'] = 'med'
         opt['alpha'] = .3
@@ -698,8 +698,8 @@ class LplsOptions(Options):
                             (blmplots.LplsHypoidCorrelationPlot, 'Hypoid corr.', False)
                             ]
         
-        opt['out_data'] = ['T','P']
+        opt['out_data'] = ['T','P', 'tsqx']
-        opt['out_plots'] = [blmplots.PlsScorePlot,blmplots.PlsLoadingPlot,blmplots.LineViewXc]
+        opt['out_plots'] = [blmplots.PlsScorePlot,blmplots.LplsXLoadingPlot,blmplots.LplsZLoadingPlot, blmplots.LineViewXc]
         
         #opt['out_data'] = None
                             

fluents/lib/blmplots.py

@@ -190,6 +190,15 @@ class PlsCorrelationLoadingPlot(BlmScatterPlot):
         title = "Pls correlation loadings (%s)" %model._dataset['X'].get_name()
         BlmScatterPlot.__init__(self, title, model, absi, ordi, part_name='CP')
         
+class LplsXLoadingPlot(BlmScatterPlot):
+    def __init__(self, model, absi=0, ordi=1):
+        title = "Lpls x-loadings (%s)" %model._dataset['X'].get_name()
+        BlmScatterPlot.__init__(self, title, model, absi, ordi, part_name='P', color_by='tsqx')
+
+class LplsZLoadingPlot(BlmScatterPlot):
+    def __init__(self, model, absi=0, ordi=1):
+        title = "Lpls z-loadings (%s)" %model._dataset['Z'].get_name()
+        BlmScatterPlot.__init__(self, title, model, absi, ordi, part_name='L', color_by='tsqz')
 
 class LplsHypoidCorrelationPlot(BlmScatterPlot):
     def __init__(self, model, absi=0, ordi=1):

fluents/lib/engines.py

@@ -219,7 +219,7 @@ def bridge(a, b, aopt, scale='scores', mode='normal', r=0):
     
 
 
-def nipals_lpls(X, Y, Z, amax, alpha=.7, mean_ctr=[2, 0, 1], mode='normal', scale='scores', verbose=False):
+def nipals_lpls(X, Y, Z, a_max, alpha=.7, mean_ctr=[2, 0, 1], mode='normal', scale='scores', verbose=False):
     """ L-shaped Partial Least Sqaures Regression by the nipals algorithm.
 
     (X!Z)->Y
@@ -260,18 +260,18 @@ def nipals_lpls(X, Y, Z, amax, alpha=.7, mean_ctr=[2, 0, 1], mode='normal', scal
     u, o = Z.shape
 
     # initialize 
-    U = empty((k, amax))
+    U = empty((k, a_max))
-    Q = empty((l, amax))
+    Q = empty((l, a_max))
-    T = empty((m, amax))
+    T = empty((m, a_max))
-    W = empty((n, amax))
+    W = empty((n, a_max))
-    P = empty((n, amax))
+    P = empty((n, a_max))
-    K = empty((o, amax))
+    K = empty((o, a_max))
-    L = empty((u, amax))
+    L = empty((u, a_max))
-    var_x = empty((amax,))
+    var_x = empty((a_max,))
-    var_y = empty((amax,))
+    var_y = empty((a_max,))
-    var_z = empty((amax,))
+    var_z = empty((a_max,))
     
-    for a in range(amax):
+    for a in range(a_max):
         if verbose:
             print "\n Working on comp. %s" %a
         u = Y[:,:1]

fluents/lib/validation.py

@@ -6,7 +6,7 @@ from scipy.stats import median
 from scipy.linalg import triu,inv,svd,norm
 
 from select_generators import w_pls_gen,w_pls_gen_jk,pls_gen,pca_gen,diag_pert
-from engines import w_simpls,pls,bridge,pca
+from engines import w_simpls,pls,bridge,pca,nipals_lpls
 from cx_utils import m_shape
 
 def w_pls_cv_val(X, Y, amax, n_blocks=None, algo='simpls'):
@@ -111,15 +111,17 @@ def pls_val(X, Y, amax=2, n_blocks=10, algo='pls', metric=None):
 
 def lpls_val(X, Y, Z, a_max=2, nsets=None,alpha=.5):
     """Performs crossvalidation to get generalisation error in lpls"""
-    cv_iter = select_generators.pls_gen(X, Y, n_blocks=nsets,center=False,index_out=True)
+    cv_iter = pls_gen(X, Y, n_blocks=nsets,center=False,index_out=True)
     k, l = Y.shape
     Yhat = empty((a_max,k,l), 'd')
     for i, (xcal,xi,ycal,yi,ind) in enumerate(cv_iter):
-        T, W, P, Q, U, L, K, B, b0, evx, evy, evz = nipals_lpls(xcal,ycal,Z,
+        dat = nipals_lpls(xcal,ycal,Z,
-                                                                a_max=a_max,
+                          a_max=a_max,
-                                                                alpha=alpha,
+                          alpha=alpha,
-                                                                mean_ctr=[2,0,1],
+                          mean_ctr=[2,0,1],
-                                                                verbose=False)
+                          verbose=False)
+        B = dat['B']
+        b0 = dat['b0']
         for a in range(a_max):
             Yhat[a,ind,:] = b0[a][0][0] + dot(xi, B[a])
     Yhat_class = zeros_like(Yhat)
@@ -135,9 +137,6 @@ def lpls_val(X, Y, Z, a_max=2, nsets=None,alpha=.5):
 def pca_alter_val(a, amax, n_sets=10, method='diag'):
     """Pca validation by altering elements in X.
 
-    
-
-
     comments:
              -- may do all jk estimates in this loop
     """
@@ -250,7 +249,7 @@ def pca_jkP(a, aopt, n_blocks=None, metric=None):
 
 
 def lpls_jk(X, Y, Z, a_max, nsets=None, alpha=.5):
-    cv_iter = select_generators.pls_gen(X, Y, n_blocks=nsets,center=False,index_out=False)
+    cv_iter = pls_gen(X, Y, n_blocks=nsets,center=False,index_out=False)
     m, n = X.shape
     k, l = Y.shape
     o, p = Z.shape
@@ -260,15 +259,11 @@ def lpls_jk(X, Y, Z, a_max, nsets=None, alpha=.5):
     WWz = empty((nsets, o, a_max), 'd')
     #WWy = empty((nsets, l, a_max), 'd')
     for i, (xcal,xi,ycal,yi) in enumerate(cv_iter):
-        T, W, P, Q, U, L, K, B, b0, evx, evy, evz = nipals_lpls(xcal,ycal,Z,
+        dat = nipals_lpls(xcal,ycal,Z,a_max=a_max,alpha=alpha,
-                                                                a_max=a_max,
+                          mean_ctr=[2,0,1],scale='loads',verbose=False)
-                                                                alpha=alpha,
+        WWx[i,:,:] = dat['W']
-                                                                mean_ctr=[2,0,1],
+        WWz[i,:,:] = dat['L']
-                                                                scale='loads',
+        #WWy[i,:,:] = dat['Q']
-                                                                verbose=False)
-        WWx[i,:,:] = W
-        WWz[i,:,:] = L
-        #WWy[i,:,:] = Q
 
     return WWx, WWz
 

scripts/yeast_annot.py

@@ -1,44 +0,0 @@
-
-def smdb_annot(orflist=None, input_fname='registry.genenames.tab', output_fname='yeast.annot'):
-
-    """Reads registry.genenames.tab from the Stanford yeast
-    microarray database.
-
-    Available from:
-    ftp://genome-ftp.stanford.edu/pub/yeast/data_download/gene_registry/registry.genenames.tab
-    
-    input: orf -- list of orfs (open reading frames)
-           file -- (optional) file to fetch info from
-           
-    registry.genames contains:
-
-    0 = Locus name 
-    1 = Other name
-    2 = Description 
-    3 = Gene product
-    4 = Phenotype
-    5 = ORF name
-    6 = SGDID
-    
-    NB! Other name, Gene product and Phenotype may have more
-    than one mapping. These are separated by |
-
-    Output: writes an annotation file
-
-    """
-    outfile = open(output_fname, 'w')
-    header = "Orf\tLocus_id\tOther_name\tDescription\tGene_product\tPhenotype\tSGD_ID\n"
-    outfile.write(header)
-    text = open(input_fname, 'r').read().splitlines()
-    for line in text:
-        els = line.split('\t')
-        orf_name = els.pop(5)
-        if orf_name!='': # we dont care about non-named orfs
-            if orflist and orf_name not in orflist:
-                break
-            for e in els:
-                if e !='':
-                    outfile.write(str(e) + "\t")
-                else:
-                    outfile.write("NA")
-            f.write("\n")