Fixed conflicts

pyblm/__init__.py

@@ -13,4 +13,3 @@ def test(level=1, verbosity=1):
     print 'Python version %s' % (sys.version.replace('\n', '',),)
     from numpy.testing import NumpyTest
     return NumpyTest().test(level, verbosity)
-

pyblm/crossvalidation.py

@@ -217,6 +217,12 @@ def lpls_val(X, Y, Z, a_max=2, nsets=None,alpha=.5, center_axis=[2,0,2], zorth=F
         # predictions
         for a in range(a_max):
             Yhat[a,val,:] = atleast_2d(ym + dot(xi, dat['B'][a]))
+    # todo: need a better support for classification error
+    y_is_class = Y.dtype.char.lower() in ['i','p', 'b', 'h','?']
+    if y_is_class:
+        pass
+        #Yhat, err = class_error(Yhat, Y)
+        #return Yhat, err
     sep = (Y - Yhat)**2
     rmsep = sqrt(sep.mean(1)).T
     #aopt = find_aopt_from_sep(rmsep)
@@ -510,6 +516,8 @@ def diag_cv(shape, nsets=9, randomise=True):
     except:
         raise ValueError("shape needs to be a two-tuple")
     if nsets>m or nsets>n:
+        msg = "You may not use more subsets than max(n_rows, n_cols)"
+        raise ValueError, msg
          msg = "You may not use more subsets than max(n_rows, n_cols)"
          nsets = min(m, n)
     nm = n*m
@@ -525,6 +533,19 @@ def diag_cv(shape, nsets=9, randomise=True):
         #training = [j for j in index if j not in validation]
         yield list(validation)
 
+def class_error(y_hat, y, method='vanilla'):
+    """ Not used.
+    """
+    a_opt, k, l = y_hat.shape
+    y_hat_c = zeros((k, l), dtype='d')
+    if method == vanilla:
+        pass
+    for a in range(a_opt):
+        for i in range(k):
+            y_hat_c[a, val, argmax(y_hat[a,val,:])] = 1.0
+    err = 100*((y_hat_c + y) == 2).sum(1)/y.sum(0).astype('d')
+
+    return y_hat_c, err
 
 def prediction_error(y_hat, y, method='squared'):
     """Loss function on multiclass Y.

pyblm/engines.py

@@ -710,13 +710,13 @@ def center(a, axis):
 
     if axis == -1:
         mn = zeros((1,a.shape[1],))
-        mn = tile(mn, (a.shape[0], 1))
+        #mn = tile(mn, (a.shape[0], 1))
     elif axis == 0:
         mn = a.mean(0)[newaxis]
-        mn = tile(mn, (a.shape[0], 1)) 
+        #mn = tile(mn, (a.shape[0], 1))
     elif axis == 1:
         mn = a.mean(1)[:,newaxis]
-        mn = tile(mn, (1, a.shape[1]))
+        #mn = tile(mn, (1, a.shape[1]))
     elif axis == 2:
         #fixme: double centering returns column mean as loc-vector, ok?
         mn = a.mean(0)[newaxis] + a.mean(1)[:,newaxis] - a.mean()
@@ -825,13 +825,12 @@ def esvd(data, a_max=None):
     """
 
     m, n = data.shape
-    if m >= n:
+    if m > n:
         kernel = dot(data.T, data)
-        
         if a_max == None:
             a_max = n - 1
         s, v = arpack.eigen_symmetric(kernel, k=a_max, which='LM',
-                                      maxiter=200, tol=1e-5)
+                                      maxiter=500, tol=1e-7)
         s = s[::-1]
         v = v[:,::-1]
         #u, s, vt = svd(kernel)
@@ -843,7 +842,7 @@ def esvd(data, a_max=None):
         if a_max == None:
             a_max = m - 1
         s, u = arpack.eigen_symmetric(kernel, k=a_max, which='LM',
-                                      maxiter=200, tol=1e-5)
+                                      maxiter=500, tol=1e-7)
         s = s[::-1]
         u = u[:,::-1]
         #u, s, vt = svd(kernel)

pyblm/statistics.py

@@ -79,18 +79,13 @@ def hotelling(Pcv, P, p_center='median', cov_center='median',
     for i in xrange(n):
         Pi = Pcv[:,i,:] # (n_sets x amax)
         Pi_ctr = P_ctr[i,:] # (1 x amax)
-        #Pim = (Pi - Pi_ctr)*msqrt(n_sets-1)
-        #Cov_i[i] = (1./n_sets)*dot(Pim.T, Pim)
-        Pim = (Pi - Pi_ctr)
-        Cov_i[i] = dot(Pim.T, Pim)
+        Pim = (Pi - Pi_ctr)*msqrt(n_sets-1)
+        Cov_i[i] = (1./n_sets)*dot(Pim.T, Pim)
+
     if cov_center == 'median':
         Cov_p = median(Cov_i)
-    elif cov_center == 'mean':
+    else cov_center == 'mean':
         Cov_p = Cov.mean(0)
-    else:
-        print "Pooled covariance est. invalid, using median"
-        print cov_center
-        Cov_p = median(Cov_i)
     reg_cov = (1. - alpha)*Cov_i + alpha*Cov_p
     for i in xrange(n):
         Pc = P_ctr[i,:]
@@ -264,7 +259,7 @@ def lpls_qvals(X, Y, Z, aopt=None, alpha=.3, zx_alpha=.5, n_iter=20,
     Wc, Lc = lpls_jk(X, Y, Z ,aopt, zorth=zorth)
     cal_tsq_x = hotelling(Wc, dat['W'], alpha=alpha)
     cal_tsq_z = hotelling(Lc, dat['L'], alpha=alpha)
-    print "morn"
+    
     # Perturbations
     index = arange(m)
     for i in range(n_iter):
@@ -403,4 +398,3 @@ def _fdr(tsq, tsqp, loc_method=median):
     fd_rate = fp/n_signif
     fd_rate[fd_rate>1] = 1
     return fd_rate
-