def __init__(self,
spikes,
history_length=100,
knot_number=5,
order_flag=1,
kernel=False):
if type(kernel) == bool:
self.history_kernel = cs.create_splines_logspace(
history_length, knot_number, 0)
else:
self.history_kernel = kernel
self.matrix = "Not initialized!"
try:
if order_flag == 1:
#print spikes.shape
#print self.history_kernel.shape
matrix = convolve_spikes(spikes, self.history_kernel)
#print matrix.shape
self.matrix = matrix[:len(spikes)]
#print self.matrix.shape
else:
self.covariate_matrix, self.covariates, self.morder = cdm.create_design_matrix_vk(
self.history_kernel, order_flag)
#print spikes.shape
#print self.covariate_matrix.shape
matrix = scipy.signal.convolve2d(spikes, self.covariate_matrix)
#print matrix.shape
self.matrix = matrix[:len(spikes)]
#print self.matrix.shape
except:
err(spikes)
err(spikes.shape)
err(self.history_kernel)
report()
raise
def fit(self, data=None,beta=None,x= None,dm = None, nr_trials=None):
"""
Fits the model
**in_spikes** `ni.data.data.Data` instance
example::
from scipy.ndimage import gaussian_filter
import ni
model = ni.model.ip.Model(ni.model.ip.Configuration({'crosshistory':False}))
data = ni.data.monkey.Data()
data = data.condition(0).trial(range(int(data.nr_trials/2)))
dm = model.dm(data)
x = model.x(data)
from sklearn import linear_model
betas = []
fm = model.fit(data)
betas.append(fm.beta)
print "fitted."
for clf in [linear_model.LinearRegression(), linear_model.RidgeCV(alphas=[0.1, 1.0, 10.0])]:
clf.fit(dm,x)
betas.append(clf.coef_)
figure()
plot(clf.coef_.transpose(),'.')
title('coefficients')
prediction = np.dot(dm,clf.coef_.transpose())
figure()
plot(prediction)
title('prediction')
ll = x * log(prediction) + (len(x)-x)*log(1-prediction)
figure()
plot(ll)
title('ll')
print np.sum(ll)
"""
fittedmodel = FittedModel(self)
fittedmodel.configuration = copy(self.configuration)
fittedmodel.history_kernel = self.history_kernel
if fittedmodel.history_kernel == "log kernel":
fittedmodel.history_kernel = cs.create_splines_logspace(self.configuration.history_length, self.configuration.knot_number, 0)
spike_train_all_trial = []
if data is not None:
spike_train_all_trial = data.getFlattend()
firing_rate = np.mean(spike_train_all_trial)
fittedmodel.firing_rate = firing_rate
fittedmodel.trial_length = data.time_bins
fittedmodel.design = self.generateDesignMatrix(data,data.time_bins)
if dm is None:
dm = self.dm(data,fittedmodel.design)
if x is None:
x = self.x(data)
w = np.where(dm.transpose())[0]
cnt = [np.sum(w== i) for i in range(dm.shape[1])]
if sum(np.array(cnt) >= dm.shape[0]) > 0:
log("!! "+str(sum(np.array(cnt) == dm.shape[0]))+ " Components are only 0. \n"+str(sum(np.array(cnt) <= dm.shape[0]*0.1))+" are mostly 0. "+str(sum(np.array(cnt) <= dm.shape[0]*0.5))+" are half 0.")
else:
log(""+str(sum(np.array(cnt) == dm.shape[0]))+ " Components are only 0. \n"+str(sum(np.array(cnt) <= dm.shape[0]*0.1))+" are mostly 0. "+str(sum(np.array(cnt) <= dm.shape[0]*0.5))+" are half 0.")
zeroed_components = [i for i in range(len(cnt)) if cnt[i] == 0]
backend_config = self.configuration.backend_config
fittedmodel.backend_model = self.backend.Model(backend_config)
if beta is None:
fit = fittedmodel.backend_model.fit(x, dm) # NOTE frequently produces LinAlgError: SVD did not converge
if False:
# This might be introduced at a later data:
# Components that are 0 can be excluded from the fitting process and then virtually reinserted in the beta and params attributes.
beta = fit.params
i_z = 0
for i in range(len(cnt)):
if i in zeroed_components:
beta.append(0)
else:
beta.append(fit.params[i_z])
i_z = i_z + 1
fit.params = beta
else:
for z in zeroed_components:
fit.params[z] = 0
fittedmodel.beta = fit.params
else:
fit = self.backend.Fit(f=None,m=fittedmodel.backend_model)
fit.params = beta
fittedmodel.beta = beta
fittedmodel.fit = fit
if "llf" in fit.statistics:
fit.statistics["llf_all"] = fit.statistics["llf"]
if hasattr(data,'nr_trials'):
fit.statistics["llf"] = fit.statistics["llf"]/data.nr_trials
elif nr_trials is not None:
fit.statistics["llf"] = fit.statistics["llf"]/nr_trials
fittedmodel.statistics = fit.statistics
#log("done fitting.")
return fittedmodel
def generateDesignMatrix(self,data,trial_length):
"""
generates a design matrix template. Uses meta data from `data` to determine number of trials and trial length.
"""
design_template = designmatrix.DesignMatrixTemplate(data.nr_trials * data.time_bins,trial_length)
log_kernel = cs.create_splines_logspace(self.configuration.history_length, self.configuration.knot_number, self.configuration.delete_last_spline)
if self.configuration.autohistory:
kernel = self.history_kernel
if kernel == "log kernel":
kernel = log_kernel
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'autohistory':
kernel = c['Kernel']
design_template.add(designmatrix.HistoryComponent('autohistory', channel=self.configuration.cell, kernel=kernel, delete_last_spline=self.configuration.delete_last_spline))
if self.configuration.autohistory_2d:
kernel_1 = self.history_kernel
kernel_2 = self.history_kernel
if kernel_1 == "log kernel":
kernel_1 = log_kernel
if kernel_2 == "log kernel":
kernel_2 = log_kernel
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'autohistory2d':
kernel_1 = c['Kernel']
kernel_2 = c['Kernel']
if c['Name'] == 'autohistory2d_1':
kernel_1 = c['Kernel']
if c['Name'] == 'autohistory2d_2':
kernel_2 = c['Kernel']
design_template.add(designmatrix.SecondOrderHistoryComponent('autohistory2d', channel_1=self.configuration.cell, channel_2=self.configuration.cell, kernel_1=kernel_1, kernel_2=kernel_2, delete_last_spline=self.configuration.delete_last_spline))
# Generating crosshistory splines for all trials
crosshistories = []
if type(self.configuration.crosshistory) == int or type(self.configuration.crosshistory) == float:
self.configuration.crosshistory = [int(self.configuration.crosshistory)]
if self.configuration.crosshistory == True or type(self.configuration.crosshistory) == list:
for i in range(data.nr_cells):
if i == self.configuration.cell:
continue
if self.configuration.crosshistory == True or i in self.configuration.crosshistory:
kernel = self.history_kernel
if kernel == "log kernel":
kernel = log_kernel
for c in self.configuration.custom_kernels:
if c['Name'] == 'crosshistory'+str(i):
kernel = c['Kernel']
design_template.add(designmatrix.HistoryComponent('crosshistory'+str(i), channel=i, kernel = kernel, delete_last_spline=self.configuration.delete_last_spline))
if self.configuration.rate:
added_rate = False
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'rate':
design_template.add(designmatrix.RateComponent('rate',kernel = c['Kernel']))
added_rate = True
if not added_rate:
if self.configuration.adaptive_rate:
rate = data.firing_rate(self.configuration.adaptive_rate_smooth_width)
design_template.add(designmatrix.AdaptiveRateComponent('rate',rate=rate,exponent=self.configuration.adaptive_rate_exponent,knots=self.configuration.knots_rate,length=trial_length))
else:
design_template.add(designmatrix.RateComponent('rate',self.configuration.knots_rate,trial_length))
if self.configuration.constant:
design_template.add(designmatrix.Component('constant',np.ones((1,1))))
##log("Combining Design Matrix")
for c in self.configuration.custom_components:
design_template.add(c)
design = design_template
design.setMask(self.configuration.mask)
return design
def addLogSpline(self, knots, header, length=0):
if length == 0:
self.add(cs.create_splines_logspace(self.length - 1, knots, 0),
header)
else:
self.add(cs.create_splines_logspace(length - 1, knots, 0), header)
def fit(self, data=None,beta=None,x= None,dm = None, nr_trials=None):
"""
Fits the model
**in_spikes** `ni.data.data.Data` instance
example::
from scipy.ndimage import gaussian_filter
import ni
model = ni.model.ip.Model(ni.model.ip.Configuration({'crosshistory':False}))
data = ni.data.monkey.Data()
data = data.condition(0).trial(range(int(data.nr_trials/2)))
dm = model.dm(data)
x = model.x(data)
from sklearn import linear_model
betas = []
fm = model.fit(data)
betas.append(fm.beta)
print "fitted."
for clf in [linear_model.LinearRegression(), linear_model.RidgeCV(alphas=[0.1, 1.0, 10.0])]:
clf.fit(dm,x)
betas.append(clf.coef_)
figure()
plot(clf.coef_.transpose(),'.')
title('coefficients')
prediction = np.dot(dm,clf.coef_.transpose())
figure()
plot(prediction)
title('prediction')
ll = x * log(prediction) + (len(x)-x)*log(1-prediction)
figure()
plot(ll)
title('ll')
print np.sum(ll)
"""
fittedmodel = FittedModel(self)
fittedmodel.configuration = copy(self.configuration)
fittedmodel.history_kernel = self.history_kernel
if fittedmodel.history_kernel == "log kernel":
fittedmodel.history_kernel = cs.create_splines_logspace(self.configuration.history_length, self.configuration.knot_number, 0)
spike_train_all_trial = []
if data is not None:
spike_train_all_trial = data.getFlattend()
firing_rate = np.mean(spike_train_all_trial)
fittedmodel.firing_rate = firing_rate
fittedmodel.trial_length = data.time_bins
fittedmodel.design = self.generateDesignMatrix(data,data.time_bins)
if dm is None:
dm = self.dm(data,fittedmodel.design)
if x is None:
x = self.x(data)
w = np.where(dm.transpose())[0]
cnt = [np.sum(w== i) for i in range(dm.shape[1])]
if sum(np.array(cnt) >= dm.shape[0]) > 0:
log("!! "+str(sum(np.array(cnt) == dm.shape[0]))+ " Components are only 0. \n"+str(sum(np.array(cnt) <= dm.shape[0]*0.1))+" are mostly 0. "+str(sum(np.array(cnt) <= dm.shape[0]*0.5))+" are half 0.")
else:
log(""+str(sum(np.array(cnt) == dm.shape[0]))+ " Components are only 0. \n"+str(sum(np.array(cnt) <= dm.shape[0]*0.1))+" are mostly 0. "+str(sum(np.array(cnt) <= dm.shape[0]*0.5))+" are half 0.")
zeroed_components = [i for i in range(len(cnt)) if cnt[i] == 0]
backend_config = self.configuration.backend_config
fittedmodel.backend_model = self.backend.Model(backend_config)
if beta is None:
fit = fittedmodel.backend_model.fit(x, dm) # NOTE frequently produces LinAlgError: SVD did not converge
if False:
# This might be introduced at a later data:
# Components that are 0 can be excluded from the fitting process and then virtually reinserted in the beta and params attributes.
beta = fit.params
i_z = 0
for i in range(len(cnt)):
if i in zeroed_components:
beta.append(0)
else:
beta.append(fit.params[i_z])
i_z = i_z + 1
fit.params = beta
else:
for z in zeroed_components:
fit.params[z] = 0
fittedmodel.beta = fit.params
else:
fit = self.backend.Fit(f=None,m=fittedmodel.backend_model)
fit.params = beta
fittedmodel.beta = beta
fittedmodel.fit = fit
if "llf" in fit.statistics:
fit.statistics["llf_all"] = fit.statistics["llf"]
if hasattr(data,'nr_trials'):
fit.statistics["llf"] = fit.statistics["llf"]/data.nr_trials
elif nr_trials is not None:
fit.statistics["llf"] = fit.statistics["llf"]/nr_trials
fittedmodel.statistics = fit.statistics
#log("done fitting.")
return fittedmodel
def generateDesignMatrix(self,data,trial_length):
"""
generates a design matrix template. Uses meta data from `data` to determine number of trials and trial length.
"""
design_template = designmatrix.DesignMatrixTemplate(data.nr_trials * data.time_bins,trial_length)
log_kernel = cs.create_splines_logspace(self.configuration.history_length, self.configuration.knot_number, self.configuration.delete_last_spline)
if self.configuration.autohistory:
kernel = self.history_kernel
if kernel == "log kernel":
kernel = log_kernel
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'autohistory':
kernel = c['Kernel']
design_template.add(designmatrix.HistoryComponent('autohistory', channel=self.configuration.cell, kernel=kernel, delete_last_spline=self.configuration.delete_last_spline))
if self.configuration.autohistory_2d:
kernel_1 = self.history_kernel
kernel_2 = self.history_kernel
if kernel_1 == "log kernel":
kernel_1 = log_kernel
if kernel_2 == "log kernel":
kernel_2 = log_kernel
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'autohistory2d':
kernel_1 = c['Kernel']
kernel_2 = c['Kernel']
if c['Name'] == 'autohistory2d_1':
kernel_1 = c['Kernel']
if c['Name'] == 'autohistory2d_2':
kernel_2 = c['Kernel']
design_template.add(designmatrix.SecondOrderHistoryComponent('autohistory2d', channel_1=self.configuration.cell, channel_2=self.configuration.cell, kernel_1=kernel_1, kernel_2=kernel_2, delete_last_spline=self.configuration.delete_last_spline))
# Generating crosshistory splines for all trials
crosshistories = []
if type(self.configuration.crosshistory) == int or type(self.configuration.crosshistory) == float:
self.configuration.crosshistory = [int(self.configuration.crosshistory)]
if self.configuration.crosshistory == True or type(self.configuration.crosshistory) == list:
for i in range(data.nr_cells):
if i == self.configuration.cell:
continue
if self.configuration.crosshistory == True or i in self.configuration.crosshistory:
kernel = self.history_kernel
if kernel == "log kernel":
kernel = log_kernel
for c in self.configuration.custom_kernels:
if c['Name'] == 'crosshistory'+str(i):
kernel = c['Kernel']
design_template.add(designmatrix.HistoryComponent('crosshistory'+str(i), channel=i, kernel = kernel, delete_last_spline=self.configuration.delete_last_spline))
if self.configuration.rate:
added_rate = False
if type(self.configuration.custom_kernels) == list:
for c in self.configuration.custom_kernels:
if c['Name'] == 'rate':
design_template.add(designmatrix.RateComponent('rate',kernel = c['Kernel']))
added_rate = True
if not added_rate:
if self.configuration.adaptive_rate:
rate = data.firing_rate(self.configuration.adaptive_rate_smooth_width)
design_template.add(designmatrix.AdaptiveRateComponent('rate',rate=rate,exponent=self.configuration.adaptive_rate_exponent,knots=self.configuration.knots_rate,length=trial_length))
else:
design_template.add(designmatrix.RateComponent('rate',self.configuration.knots_rate,trial_length))
if self.configuration.constant:
design_template.add(designmatrix.Component('constant',np.ones((1,1))))
##log("Combining Design Matrix")
for c in self.configuration.custom_components:
design_template.add(c)
design = design_template
design.setMask(self.configuration.mask)
return design
