Projects/laydi
Archived
7
0
Fork 0
Code Issues Pull Requests Activity

... just lots of stuff

Browse Source
This commit is contained in:
Arnar Flatberg
2007-08-24 09:14:24 +00:00
parent 21b63b17e5
commit 7dbf28f65d
6 changed files with 134 additions and 109 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fluents/lib/blmfuncs.py
+59 -50
View File
@@ -18,6 +18,39 @@ from packer import Packer
import blmplots
class NewStyleModel(Function):
def __init__(self, id='johndoe', name='JohnDoe'):
Function.__init__(self, id, name)
self.name = name
self.options = Options
self.input_data = []
self.parts = {}
self.io_table = {}
self.datasets = []
self.plots = []
def create_dataset(self, param, Dataset=Dataset):
for ds in self.datasets:
if ds.get_name()==param: return ds
if not param in self.parts.keys():
logger.log('notice', 'Parameter: %s not present' %param)
return
if not param in self.io_table.keys():
logger.log('notice', 'Parameter: %s not in defined in io table' %param)
identifiers = self.io_table.get(param)
data = self.parts.get(param)
ds = Dataset(data, identifiers=identifiers, name=param)
self.datasets.append(dataset)
return ds
def create_plot(self, blmplot):
if blmplot.validate_model(self.parts):
plt = blmplot(self.parts)
self.plots.append(plt)
return plt
class Model(Function):
"""Base class of bilinear models.
"""
@@ -68,7 +101,7 @@ class PCA(Model):
logger.log('notice', 'Aopt at first component!')
def confidence(self, aopt, n_sets, alpha, p_center,
crot, strict, cov_center ):
crot, strict, cov_center):
"""Returns a confidence measure for model parameters.
Based on aopt.
@@ -80,7 +113,7 @@ class PCA(Model):
jk_segments = pca_jkP(self.model['E0'], aopt, n_sets)
Pcal = self.model['P'][:,:aopt]
# add the scale to P
# ensure scaled P
tnorm = scipy.apply_along_axis(norm, 0, self.model['T'][:,:aopt])
Pcal = Pcal*tnorm
tsq = hotelling(jk_segments, Pcal, p_center,
@@ -90,33 +123,14 @@ class PCA(Model):
def make_model(self, amax, mode, scale):
"""Model on optimal number of components.
"""
dat = pca(self.model['E0'], amax, scale, mode)
# explained variance
var_x, exp_var_x = variances(self.model['E0'], dat['T'], dat['P'])
dat['var_x'] = var_x
dat['exp_var_x'] = exp_var_x
#fixme###
do_lev_s = False
do_lev_v = False
#####
if do_lev_s:
# sample leverages
tnorm = scipy.apply_along_axis(norm, 0, dat['T']) # norm of Ts
s_lev = leverage(amax, tnorm)
dat['s_lev'] = s_lev
if do_lev_v:
# variable leverages
v_lev = leverage(amax, dat['P'])
dat['v_lev'] = v_lev
dat = pca(self.model['E0'], amax, scale, mode)
self.model.update(dat)
def as_dataset(self, param, dtype='dataset'):
"""Return model parameter as Dataset.
"""
if not param in self.model.keys():
logger.log('notice', 'Parameter: %s not in model' %param)
return
DX = self._dataset['X'] #input dataset
dim_name_0, dim_name_1 = DX.get_dim_name()
@@ -128,17 +142,18 @@ class PCA(Model):
pc_ids = ['_amax', map(str,range(self._options['amax'])) ]
pc_ids_opt = ['_aopt', map(str, range(self._options['aopt'])) ]
zero_dim = ['_doe', ['0']] # null dim, vector (hidden)
match_ids = {'E':[ids_0, ids_1],
'E0':[ids_0, ids_1],
'P':[ids_1, pc_ids],
'T':[ids_0, pc_ids],
'W':[ids_1, pc_ids],
'p_tsq':[ids_1, zero_dim],
'rmsep':[pc_ids, zero_dim],
'var_leverages':[ids_1, zero_dim],
'sample_leverages':[pc_ids, zero_dim],
'exp_var_x': [pc_ids, zero_dim],
'var_x': [pc_ids, zero_dim],
match_ids = {'E' : [ids_0, ids_1],
'E0' : [ids_0, ids_1],
'P' : [ids_1, pc_ids],
'T' : [ids_0, pc_ids],
'W' : [ids_1, pc_ids],
'p_tsq' : [ids_1, zero_dim],
'rmsep' : [pc_ids, zero_dim],
'var_leverages' : [ids_1, zero_dim],
'sample_leverages' : [pc_ids, zero_dim],
'exp_var_x' : [pc_ids, zero_dim],
'var_x' : [pc_ids, zero_dim],
'eigvals' : [pc_ids, zero_dim]
}
out = Dataset(self.model[param], match_ids[param], name=param)
@@ -256,7 +271,7 @@ class PLS(Model):
def make_model(self, a, b, amax, scale, mode, engine):
"""Make model on amax components.
"""
print "MAking model"
print "Making model"
dat = engine(a, b, amax, scale, mode)
self.model.update(dat)
@@ -478,13 +493,6 @@ class LPLS(Model):
self._data['Z'] = c.asarray()
self.validation(options)
self.make_model(options)
print self.model['evx']
evx_str = [str(i)[:3] for i in self.model['evx']]
logger.log('notice', 'Explained variance:X\n\t: ' + str(evx_str))
evy_str = [str(i)[:3] for i in self.model['evy']]
logger.log('notice', 'Explained variance:Y\n\t: ' + str(evy_str))
evz_str = [str(i)[:3] for i in self.model['evz']]
logger.log('notice', 'Explained variance:Z\n\t: ' + str(evz_str))
if options['calc_conf']:
self.confidence(options)
@@ -553,13 +561,15 @@ class PcaOptions(Options):
opt['all_plots'] = [(blmplots.PcaScorePlot, 'Scores', True),
(blmplots.PcaLoadingPlot, 'Loadings', True),
(blmplots.LineViewXc, 'Line view', True),
(blmplots.PredictionErrorPlot, 'Residual Error', False)
(blmplots.PredictionErrorPlot, 'Residual Error', False),
(blmplots.PcaScreePlot, 'Scree', True)
]
opt['out_data'] = ['T','P', 'p_tsq']
opt['out_plots'] = [blmplots.PcaScorePlot,
blmplots.PcaLoadingPlot,
blmplots.LineViewXc]
blmplots.LineViewXc,
blmplots.PcaScreePlot]
self.update(opt)
@@ -621,7 +631,8 @@ class PlsOptions(Options):
# (class, name, sensitive, ticked)
opt['all_plots'] = [(blmplots.PlsScorePlot, 'Scores', True),
(blmplots.PlsLoadingPlot, 'Loadings', True),
(blmplots.PlsXLoadingPlot, 'X-Loadings', True),
(blmplots.PlsYLoadingPlot, 'Y-Loadings', True),
(blmplots.LineViewXc, 'Line view', True),
(blmplots.PredictionErrorPlot, 'Residual Error', False),
(blmplots.RMSEPPlot, 'RMSEP', False),
@@ -629,7 +640,7 @@ class PlsOptions(Options):
]
opt['out_data'] = ['T','P', 'w_tsq']
opt['out_plots'] = [blmplots.PlsScorePlot,blmplots.PlsLoadingPlot,blmplots.LineViewXc]
opt['out_plots'] = [blmplots.PlsScorePlot,blmplots.PlsXLoadingPlot,blmplots.PlsYLoadingPlot,blmplots.LineViewXc]
#opt['out_data'] = None
@@ -699,11 +710,9 @@ class LplsOptions(Options):
]
# (class, name, sensitive, ticked)
opt['all_plots'] = [(blmplots.PlsScorePlot, 'Scores', True),
(blmplots.PlsLoadingPlot, 'Loadings', True),
opt['all_plots'] = [(blmplots.LplsScorePlot, 'Scores', True),
(blmplots.LplsXLoadingPlot, 'Loadings', True),
(blmplots.LineViewXc, 'Line view', True),
(blmplots.PredictionErrorPlot, 'Residual Error', False),
(blmplots.RMSEPPlot, 'RMSEP', False),
(blmplots.LplsHypoidCorrelationPlot, 'Hypoid corr.', False),
(blmplots.LplsXCorrelationPlot, 'X corr.', True),
(blmplots.LplsZCorrelationPlot, 'Z corr.', True)