class DSMMeasure(Measure):
"""DSMMeasure creates a Measure object
where metric can be one of 'euclidean', 'spearman', 'pearson'
or 'confusion'"""
def __init__(self, dsmatrix, dset_metric, output_metric='spearman'):
Measure.__init__(self)
self.dsmatrix = dsmatrix
self.dset_metric = dset_metric
self.output_metric = output_metric
self.dset_dsm = []
def __call__(self, dataset):
# create the dissimilarity matrix for the data in the input dataset
self.dset_dsm = DSMatrix(dataset.samples, self.dset_metric)
in_vec = self.dsmatrix.get_vector_form()
dset_vec = self.dset_dsm.get_vector_form()
# concatenate the two vectors, send to dissimlarity function
test_mat = np.asarray([in_vec, dset_vec])
test_dsmatrix = DSMatrix(test_mat, self.output_metric)
# return correct dissimilarity value
return test_dsmatrix.get_full_matrix()[0, 1]
class DSMMeasure(Measure):
"""DSMMeasure creates a Measure object
where metric can be one of 'euclidean', 'spearman', 'pearson'
or 'confusion'"""
def __init__(self, dsmatrix, dset_metric, output_metric="spearman"):
Measure.__init__(self)
self.dsmatrix = dsmatrix
self.dset_metric = dset_metric
self.output_metric = output_metric
self.dset_dsm = []
def __call__(self, dataset):
# create the dissimilarity matrix for the data in the input dataset
self.dset_dsm = DSMatrix(dataset.samples, self.dset_metric)
in_vec = self.dsmatrix.get_vector_form()
dset_vec = self.dset_dsm.get_vector_form()
# concatenate the two vectors, send to dissimlarity function
test_mat = np.asarray([in_vec, dset_vec])
test_dsmatrix = DSMatrix(test_mat, self.output_metric)
# return correct dissimilarity value
return test_dsmatrix.get_full_matrix()[0, 1]
def _call(self, dataset):
# compute rsm
dsm = DSMatrix(dataset.samples, metric=self.metric)
vec_form = samples = dsm.get_vector_form()
n = dsm.full_matrix.shape[0]
# mask for upper diagonal
msk = np.asarray([[i > j for i in xrange(n)] for j in xrange(n)])
# compute z scores
arr = np.reshape(np.asarray(dsm.full_matrix[msk]), (-1, ))
arr_z = (arr - np.mean(arr)) / np.std(arr)
# set the space
ds = Dataset(samples=arr_z)
if not self.space is None:
space_vals = dataset.sa[self.space].value
labels = [
'%s-%s' % (space_vals[i], space_vals[j]) for i in xrange(n)
for j in xrange(n) if msk[i, j]
]
ds.sa[self.space] = labels
return ds
def _call(self, dataset):
# compute rsm
dsm = DSMatrix(dataset.samples, metric=self.metric)
vec_form = samples = dsm.get_vector_form()
n = dsm.full_matrix.shape[0]
# mask for upper diagonal
msk = np.asarray([[i > j for i in xrange(n)] for j in xrange(n)])
# compute z scores
arr = np.reshape(np.asarray(dsm.full_matrix[msk]), (-1,))
arr_z = (arr - np.mean(arr)) / np.std(arr)
# set the space
ds = Dataset(samples=arr_z)
if not self.space is None:
space_vals = dataset.sa[self.space].value
labels = ['%s-%s' % (space_vals[i], space_vals[j])
for i in xrange(n) for j in xrange(n) if msk[i, j]]
ds.sa[self.space] = labels
return ds
