def main():
# data = hddm.load_csv('/data/pdmattention/TestData.csv')
data = hddm.load_csv('/data/pdmattention/TestData.csv')
data.head(10) # returns first n rows for the object based on position
# data = hddm.utils.flip_errors(data) # flips sign for lower boundary
# response --> what are lower
# boundaries
# fig = plt.figure() # creates an empty figure
# ax = fig.add_subplot(111, xlabel = 'RT', ylabel='count', title='RT distributions')
# for i, subj_data in data.groupby('subj_idx'):
# data.groupby splits that data into the groups based on some criterion
# grouped = obj.groupby(key)
# grouped = obj.groupby(key, axis = 1)
# grouped = obj.groupby([key1, key2])
# data.groupby('subj_idx') groups data horizontally by
# the column subj_idx
# for i, subj_data in data.groupby('subj_idx'):
# subj_data.rt.hist(bins=20, histtype='step', ax=ax)
# plt.savefig('/data/pdmattention/hddm_fig1.pdf')
# plt.show()
# time.sleep(5)
# print('\n')
# print('Done!')
next_step(data)
def run_exp1mod16(id):
""" drift-rate: rwd/pen tradeoff, stable target variance model
bias : rwd/pen tradeoff, stable target variance model"""
#imports
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp1data.csv')
#data grooming, remove fast outliers
rtSig = data.rt.std()
rtMu = data.rt.mean()
cutoff = rtMu - rtSig
data = data[data.rt > cutoff]
data.reset_index(drop=True, inplace=True)
#build model
exp1model16 = hddm.HDDM(data,
depends_on={
'v': ['rwd_pen', 'stableVar'],
'z': ['rwd_pen', 'stabelVar']
},
bias=True,
include=['v', 'a', 't', 'z'],
p_outlier=0.05)
exp1model16.find_starting_values()
exp1model16.sample(3000,
burn=1000,
dbname='exp1mod16_%i.db' % id,
db='pickle')
return exp1model16
def main():
path = "/home/megan/Desktop/Drift-Diffusion/data"
data = hddm.load_csv(
"/home/megan/Desktop/Drift-Diffusion/modeling/diffusion-nov-cutoff.csv"
)
#visualizeRT(data)
#simpleModel(data)
#GelmanRubinModel(data)
modelFitDrift(data)
def setUp(self):
print("Passing setUp")
self.models = list(
hddm.torch.torch_config.TorchConfig(model="ddm").network_files.
keys()) + ["ddm_vanilla", "full_ddm_vanilla"]
self.n_samples_per_subject = 100
self.n_trials = 10
self.n_subjects = 5
self.cav_data = hddm.load_csv(hddm.__path__[0] +
"/examples/cavanagh_theta_nn.csv")
pass
def run_testmod(id):
#imports
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp1data.csv')
#data grooming, remove fast outliers
rtSig = data.rt.std()
rtMu = data.rt.mean()
cutoff = rtMu - rtSig
data = data[data.rt > cutoff]
data.reset_index(drop=True, inplace=True)
#build model
testmod = hddm.HDDM(data)
testmod.sample(200, burn=20, dbname='testmod_%i.db' % id, db='pickle')
return testmod
def setUp(self):
self.models = list(
hddm.torch.torch_config.TorchConfig(
model="ddm").network_files.keys())
self.n_samples_per_trial = 1000
self.nmcmc = 200
self.nburn = 100
self.filepath = hddm.__path__[0] + "/tests/torch_models/"
self.cav_data = hddm.load_csv(hddm.__path__[0] +
"/examples/cavanagh_theta_nn.csv")
# Make model test folder if it is not there
if not os.path.isdir(self.filepath):
os.mkdir(self.filepath)
pass
def run_exp1mod6(id):
""" drift-rate: stable target variance model """
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp1data.csv')
rtSig = data.rt.std()
rtMu = data.rt.mean()
cutoff = rtMu - rtSig
data = data[data.rt > cutoff]
data.reset_index(drop=True, inplace=True)
exp1model6 = hddm.HDDM(data, depends_on={'v': 'stableVar'}, p_outlier=0.05)
exp1model6.find_starting_values()
exp1model6.sample(1500,
burn=500,
dbname='exp1mod6_%i.db' % id,
db='pickle')
return exp1model6
def run_exp2mod6(id):
""" drift-rate: rwd prob model """
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp2data.csv')
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
exp2model6 = hddm.HDDM(data, depends_on={'v': 'rwd_prob'}, p_outlier=0.05)
exp2model6.find_starting_values()
exp2model6.sample(1500,
burn=500,
dbname='exp2mod6_%i.db' % id,
db='pickle')
return exp2model6
def run_exp2mod2(id):
"""
Bias: task params model
"""
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp2data.csv')
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
exp2model2 = hddm.HDDM(data,
depends_on={'z': ['rwd_pen', 'rwd_prob']},
include='z',
p_outlier=0.05)
exp2model2.find_starting_values()
exp2model2.sample(1500, burn=500, dbname='exp2mod2%i.db' % id, db='pickle')
return exp2model2
def run_exp1mod2(id):
""" bias:task params model """
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp1data.csv')
rtSig = data.rt.std()
rtMu = data.rt.mean()
cutoff = rtMu - rtSig
data = data[data.rt > cutoff]
data.reset_index(drop=True, inplace=True)
exp1model2 = hddm.HDDM(data,
depends_on={'z': ['rwd_pen', 'stableVar']},
include='z',
p_outlier=0.05)
exp1model2.find_starting_values()
exp1model2.sample(1500,
burn=500,
dbname='exp1mod2_%i.db' % id,
db='pickle')
return exp1model2
def run_exp3mod1(id):
""" Pure drift-rate to task params model
"""
import hddm
data = hddm.load_csv(
'C:/Users/Rory/Dropbox/Net_Worth/Pub_Code/Data/Experiment3/Model/Exp3ModelData.csv'
)
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
exp3model1 = hddm.HDDM(data,
depends_on={'v': ['rwd_prob', 'stableVar']},
p_outlier=0.05)
exp3model1.find_starting_values()
exp3model1.sample(1500, burn=250, dbname='exp3mod1%i.db' % id, db='pickle')
return exp3model1
def gelman_rubin_models(n):
import hddm
import pandas as pd
import pickle
import os
#load data
data = hddm.load_csv(
'https://raw.githubusercontent.com/xuankai91/MResProject/master/tofit_HDDM.csv'
)
print('data loaded')
#setup
samples = 5000
savepath = './'
#create save folders
if not os.path.exists(savepath + 'gelman_rubin'):
os.makedirs(savepath + 'gelman_rubin')
#start modelling
print('starting...')
#instantiate model object
model = hddm.HDDM(data,
depends_on={
'v': 'stim',
'a': 'stim',
't': 'stim'
},
include=('sv', 'st', 'sz'),
p_outlier=.05)
#start model fitting
model.find_starting_values()
model.sample(samples,
burn=int(samples / 10),
dbname='%s/gelman_rubin/m_grs_%d_traces.db' % (savepath, n),
db='pickle')
model.save('%s/gelman_rubin/m_grs_%d' % (savepath, n))
def run_exp1mod1(id):
""" pure drift-rate:task params model """
#imports
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp1data.csv')
#data grooming, remove fast outliers
rtSig = data.rt.std()
rtMu = data.rt.mean()
cutoff = rtMu - rtSig
data = data[data.rt > cutoff]
data.reset_index(drop=True, inplace=True)
#build model
exp1model1 = hddm.HDDM(data,
depends_on={'v': ['rwd_pen', 'stableVar']},
p_outlier=0.05)
exp1model1.find_starting_values()
exp1model1.sample(3000,
burn=1500,
dbname='exp1mod1_%i.db' % id,
db='pickle')
return exp1model1
def run_exp3mod8(id):
""" bias: rwd prob model """
import hddm
data = hddm.load_csv(
'C:/Users/Rory/Dropbox/Net_Worth/Pub_Code/Data/Experiment3/Model/Exp3ModelData.csv'
)
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
exp3model8 = hddm.HDDM(data,
depends_on={'z': 'rwd_prob'},
include='z',
p_outlier=0.05)
exp3model8.find_starting_values()
exp3model8.sample(1500,
burn=250,
dbname='exp3mod8_%i.db' % id,
db='pickle')
return exp3model8
def run_exp2mod10(id):
""" drift-rate: rwd/pen tradeoff with free bias model """
#imports
import hddm
data = hddm.load_csv('~/Pub_Code_master/Data/exp2data.csv')
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
#build model
exp2model10 = hddm.HDDM(data,
depends_on={'v': 'rwd_pen'},
include='z',
p_outlier=0.05)
exp2model10.find_starting_values()
exp2model10.sample(1500,
burn=500,
dbname='exp2mod10_%i.db' % id,
db='pickle')
return exp2model10
def run_exp3mod10(id):
""" drift-rate: stable target variance with free bias model """
#imports
import hddm
data = hddm.load_csv(
'C://Users/Rory/Dropbox/Net_Worth/Pub_Code/Data/Experiment3/Model/Exp3ModelData.csv'
)
#data grooming, remove fast outliers
#rtSig = data.rt.std()
#rtMu = data.rt.mean()
#cutoff = rtMu - rtSig
#data = data[data.rt>cutoff]
#data.reset_index(drop=True, inplace=True)
#build model
exp3model10 = hddm.HDDM(data,
depends_on={'v': 'stableVar'},
include='z',
p_outlier=0.05)
exp3model10.find_starting_values()
exp3model10.sample(1500,
burn=250,
dbname='exp3mod10_%i.db' % id,
db='pickle')
return exp3model10
def run_model(id):
from patsy import dmatrix
from pandas import Series
import numpy as np
import hddm
dataHDDM = hddm.load_csv('DDM/dataHDDM_pmt.csv')
dataHDDM["subj_idx"] = dataHDDM["participant"]
del dataHDDM["participant"]
dataHDDM["SAT"] = dataHDDM.apply(lambda row: 0 if row['SAT'] == "Accuracy" else 1, axis=1)
dataHDDM["FC"] = dataHDDM.apply(lambda row: -0.5 if row['FC'] == "low" else 0.5, axis=1)
dataHDDM["contrast"] = dataHDDM.contrast.replace([1,2,3,4,5,6], [-.5,-.3,-.1,.1,.3,.5])
dataHDDM["givenResp"] = dataHDDM["response"]
dataHDDM["stim"] = dataHDDM.apply(lambda row: 1 if row['stim'] == 'Right' else 0, axis=1)
dataHDDM["response"] = dataHDDM.apply(lambda row: 1 if row['givenResp'] == 'Right' else 0, axis=1)
def v_link_func(x, data=dataHDDM):
stim = (np.asarray(dmatrix('0 + C(s, [[1], [-1]])',
{'s': data.stim.ix[x.index]})))
return x*stim
if id < 4:
############## M1
LM = [{'model':'t ~ SAT + FC + contrast + SAT:FC + SAT:contrast + FC:contrast + SAT:FC:contrast', 'link_func': lambda x: x} ,
{'model':'v ~ contrast', 'link_func':v_link_func} ,
{'model':'a ~ FC + SAT + SAT:FC', 'link_func': lambda x: x} ]
deps = {'sz' : 'SAT'}
inc = ['sv','sz','st','z']
model_name = "Joint_t0"
else :
return np.nan()
name = 'light_reg_PMT_%s' %str(id)
m = hddm.HDDMRegressor(dataHDDM, LM , depends_on = deps,
include=inc, group_only_nodes=['sv', 'sz','st', "sz_SAT"], group_only_regressors=False, keep_regressor_trace=True)
m.find_starting_values()
m.sample(iter=10000, burn=8500, thin=1, dbname='DDM/traces/db_%s'%name, db='pickle')
m.save('DDM/Fits/%s'%name)
return m
import hddm
print(hddm.__version__)
# Python version:
import sys
print (sys.version)
print (pd.__version__)
# Adding a monkey in order to save the models dbs:
import pickle
# Get around a problem with saving regression outputs in Python 3
def savePatch(self, fname):
with open(fname, 'wb') as f:
pickle.dump(self, f)
hddm.HDDM.savePatch = savePatch
data = hddm.load_csv("/ems/elsc-labs/loewenstein-y/lior.lebovich/DD/g27/data_csv_g27_dl0_3columns.csv")
# data = data[data.subj_idx!=125]
docdir = "/ems/elsc-labs/loewenstein-y/lior.lebovich/DD/g27/none1"
os.chdir(docdir)
i = sys.argv[1]
print(i)
m = hddm.HDDM(data,
p_outlier=.05, informative=True)
m.find_starting_values()
m.sample(40000, burn=20000, thin=5, dbname='db'+str(i), db='pickle')
m.savePatch('submodel'+str(i))
print('Thats it!!')
'Nburn: ' + str(nburn) +
'\n################################################\n################################################\n'
)
#if args.verbose==None:
# verb = True
#else:
# verb = False
########################
##############################################
## load data and flip incorrect RTs if accuracy coding
##############################################
print('loading raw RT data from: ' + rawdf)
data = hddm.load_csv(rawdf)
print('data structure: ')
print(data.head(5))
if code == 'acc':
print('\nflipping RT errors\n')
data = hddm.utils.flip_errors(data)
## navigate to outputdir
os.chdir(outputdir)
print('Navigated to ' + os.getcwd())
print('Directory contents:\n')
print(os.listdir('.'))
import pandas as pd
import matplotlib.pyplot as plt
import hddm
import pickle
# Load data from csv file into a NumPy structured array
data = hddm.load_csv('C://python//hddm//data//simple_hddm.csv')
# Create a HDDM model multi object
model = hddm.HDDM(data)
# find a good starting point which helps with the convergence.
model.find_starting_values()
# Create model and start MCMC sampling
model.sample(2000, burn=20, dbname='traces.db', db='pickle')
# Print fitted parameters and other model statistics
model.save('C://python//hddm//models//simple//mymodel_jupyter')
code = 'stim' #coding scheme: stimulus(stim) or accuracy(acc)
nchains = 5
# vestigial from earlier days
#nmodels=6 #number of distinct models to run below
#mods = 1 #estimate baseline models?
#SNAP_mods = 0 #estimate dimensional models
######################################################################
##Read in data and set up burn and sample quantities
if ics == 0:
#local
basedir = '/Users/natehall/github_repos/PD_Inhibition_DDM'
os.chdir(basedir)
if code == 'stim':
data = hddm.load_csv('Data/preprocessed/flank_use.csv')
elif code == 'acc':
# data = hddm.load_csv('Flanker_reduced_acc_coded_nooutliers.csv')
outputdir = basedir+'/Outputs/practice_hddm'
nsample = 50
nburn = 3
os.chdir(outputdir)
elif ics == 1:
os.chdir('/gpfs/group/mnh5174/default/DEPENd_Box/Projects/PD_Inhibition_DDM')
if code == 'stim':
data = hddm.load_csv('Flanker_reduced_stim_coded_nooutliers.csv')
elif code == 'acc':
data = hddm.load_csv('Flanker_reduced_acc_coded_nooutliers.csv')
#outputdir = ('practice_files')
nsample = 20000
nburn = 5000
os.getcwd()
os.chdir("/home/brain/host/hddm_exp7_rep/")
# get the tool box
import pandas as pd
import matplotlib.pyplot as plt
# plt.rcParams['image.cmap'] = 'viridis' # change default colormap
import hddm
import time
from cycler import cycler
plt.rcParams['axes.prop_cycle'] = cycler(color='bgrcmyk')
# Load match data from csv file into a NumPy structured array
dat_C_Id = hddm.load_csv('exp7_rep_categ_id_hddm.csv')
dat_C_Id.head(10)
# flip the error RTs to be negative
dat_C_Id = hddm.utils.flip_errors(dat_C_Id)
# check the RT distritubtion
fig = plt.figure()
ax = fig.add_subplot(111,
xlabel='RT',
ylabel='count',
title='RT distributions')
for i, subj_data in dat_C_Id.groupby('subj_idx'):
subj_data.rt.hist(bins=20, histtype='step', ax=ax)
plt.savefig('p_exp7_rep_Id_flipped.pdf')
# Run the Hierarchical Drift Diffusion Model (HDDM; Wiecki T.V, et al., 2013) on the discriminability data
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import hddm
# %matplotlib inline
# load the data
dt_data = hddm.load_csv('./hddm_dt_data.csv')
# dt data:
'''
subjects: 19 (0-18)
pairs: 666 (630 different, 36 same)
rt: reaction times (10 rts per different pair, 56 rts per same pair)
response: correct(1), incorrect(0)
condition: same objects pair(1), different objects pair(2)
'''
# add headers
headers = ["subNo", "pair", "rt", "response", "condition"]
dt_data.columns = headers
# in python indexing starts with zero, thus, subject one is actually zero. Change indexing
dt_data["subNo"] = dt_data["subNo"] - 1
# quick check
dt_data.head(20)
import pandas as pd
import hddm
import pickle
# load data
data = hddm.load_csv(
'Z://Work//UW//projects//RR_TMS//hddm//data//infins_hddm.csv')
# init models
models = []
# make models
for i in range(5):
m = hddm.HDDM(data, depends_on={'v': 'stim', 'a': 'stim'})
m.find_starting_values()
m.sample(
10000,
burn=5000,
dbname=
'Z://Work//UW//projects//RR_TMS//hddm//db//ii_va_stim_traces%i.db' % i,
db='pickle')
models.append(m)
# save models
for i in range(5):
fname = 'Z://Work//UW//projects//RR_TMS//hddm//models//by_cond//ii_va_stim' + str(
i)
models[i].save(fname)
import pandas as pd
import matplotlib.pyplot as plt
import numpy as npy
# matplotlib inline
import hddm
from patsy import dmatrix
data = hddm.load_csv('data/fmri_modeling.csv')
data = hddm.utils.flip_errors(data)
data = data[data.run > 2] # remove counts
data = data[~npy.isnan(data.rt)] #remove nans
fig = plt.figure()
ax = fig.add_subplot(111,
xlabel='RT',
ylabel='count',
title='RT distributions')
for i, subj_data in data.groupby('subj_idx'):
subj_data.rt.hist(bins=20, histtype='step', ax=ax)
plt.savefig('fMRI_rt.pdf')
#
#
dmatrix("C(stim,Treatment(1))", data.head(10))
m_stim = hddm.HDDMStimCoding(data,
include='z',
stim_col='stim',
split_param='z',
depends_on={
'v': 'response',
'a': 'response',
##Create a string for SNAP (Schedule for Nonadaptive and Adaptive Personality) dimensions
#SNAP_dim = ('DISINH', 'SUICPRON', 'NEGTEMP', 'MISTRUST', 'MANIP', 'AGG', 'SELFHARM', 'ECCPERC', 'DEPEN', 'POSTEMP', 'EXHIB', 'ENTITL', 'DETACH', 'IMPUL', 'PROPER', 'HDWK','DISINHP', 'LOSLFEST')
#nmodels=6 #number of distinct models to run below
#mods = 1 #estimate baseline models?
#SNAP_mods = 0 #estimate dimensional models
######################################################################
##Read in data and set up burn and sample quantities
if ics == 0:
#local
basedir = '/Users/natehall/github_repos/PD_Inhibition_DDM'
os.chdir(basedir)
if code == 'stim':
data = hddm.load_csv('Data/preprocessed/flank_stimCode.csv')
elif code == 'acc':
data = hddm.load_csv('Data/preprocessed/flank_accCode.csv')
outputdir = basedir+'/Outputs/practice_hddm/flanker'
nsample = 10
nburn = 3
elif ics == 1:
basedir = '/gpfs/group/mnh5174/default/Nate/PD_Inhibition_DDM'
os.chdir(basedir)
if code == 'stim':
data = hddm.load_csv('Data/preprocessed/flank_stimCode.csv')
elif code == 'acc':
data = hddm.load_csv('Data/preprocessed/flank_accCode.csv')
data = hddm.utils.flip_errors(data)
if short_chains:
outputdir = basedir+'/Outputs/practice_hddm/flanker'
import pandas as pd
import hddm
import pickle
# load data
data = hddm.load_csv(
'Z://Work//UW//projects//RR_TMS//hddm//data//fullsplit_hddm.csv')
regm = hddm.HDDMRegressor(data,
"z ~ rulert:C(tms, Treatment('NV'))",
depends_on={
'v': 'stim',
'a': 'stim'
},
bias=True,
include='all',
p_outlier=0.05)
regm.find_starting_values()
regm.sample(
2000,
burn=200,
dbname='Z://Work//UW//projects//RR_TMS//hddm//db//testmodel_traces.db',
db='pickle')
fname = 'Z://Work//UW//projects//RR_TMS//hddm//models//by_cond//testmodel'
regm.save(fname)
import hddm
# Load data from csv file into a NumPy structured array
data = hddm.load_csv('simple_difficulty.csv')
# Create a HDDM model multi object
model = hddm.HDDM(data, depends_on={'v':'difficulty'})
# Create model and start MCMC sampling
model.sample(10000, burn=5000)
# Print fitted parameters and other model statistics
model.print_stats()
# Plot posterior distributions and theoretical RT distributions
model.plot_posteriors(save=True)
model.plot_posterior_predictive(save=True)
import os
import hddm
import matplotlib.pyplot as plt
#folder = '/home/ausmanpa/Documents/gp/ACTR_DDM/simulations/'
folder = '/projects/actr/models/ACTR_DDM/simulations01_redux/'
files = os.listdir(folder)
folderIdx = files.index('DDM_results')
files.pop(folderIdx)
#listdir will list the results folder, remove it
for f in files:
hddmData = hddm.load_csv(folder + f)
#create model object where v, a, and t all depend on the difficulty
model = hddm.HDDM(hddmData,
depends_on={
'v': 'stim',
'a': 'stim',
't': 'stim'
})
#find a good starting point to help with convergence - but, seems to run into a warning/error
model.find_starting_values()
#draw 2000 samples, discard 20 as burn-in
model.sample(2000, burn=20, dbname='traces.db', db='pickle')
## Speed-Acc HDDM Test
## Aviva Blonder
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import hddm
## Loading and arranging data for analysis
data = hddm.load_csv("data/3_final_merged_data/speed-acc-ss-df.csv")
data = data.dropna(subset = ['RT'])
data['stimuli'] = np.where(data['stimuli'] == "V1", "ASL", data['stimuli'])
data['stimuli'] = np.where(data['stimuli'] == "V2", "ASL", data['stimuli'])
data['stimuli'] = np.where(data['stimuli'] == "Trio", "Object", data['stimuli'])
data['stimuli'] = np.where(data['stimuli'] == "Bull", "Bullseye", data['stimuli'])
data['rt'] = data['RT_sec']
data['response'] = data['correct']
# print(data.head(10))
#%%
## Mixture model - no parameters vary by condition
model = hddm.HDDM(data, include = ('p_outlier'))
model.find_starting_values()
model.sample(2000, burn = 20)
model.print_stats()
model.plot_posteriors()
model.plot_posterior_predictive(figsize = (14, 10))
#%%
## Drift and boundary separation can vary by condition
var_model = hddm.HDDM(data, include = ('p_outlier'), depends_on = {'v': 'stimuli', 'a' : 'stimuli'})
def run_model(trace_id):
import os, pickle, sys, time, hddm
model_name = 'basic_stimcoding'
nsamples = 10000 # 50.000 for real results?
# find path depending on local/klimag
usr = os.environ.get('USER')
if usr in ['anne']:
mypath = '/Users/anne/Data/projects/0/neurodec/Data/MEG-PL/Data/HDDM'
if usr in ['aurai']:
mypath = '/home/aurai/Data/MEG-PL/Data/HDDM'
# make a folder for the outputs, combine name and time
# currentTime = datetime.now().strftime('%Y-%m-%d-%H-%M-%S')
# thispath = os.path.join(mypath, model_name + '_' + currentTime)
thispath = os.path.join(mypath, model_name)
if not os.path.exists(thispath):
os.mkdir(thispath)
# ============================================ #
# load in data
# ============================================ #
mydata = hddm.load_csv(os.path.join(mypath, '2ifc_data_hddm.csv'))
# specify the model
m = hddm.HDDMStimCoding(mydata, stim_col='stimulus', split_param='v',
drift_criterion=True, bias=True,
include=('sv'), group_only_nodes=['sv'],
depends_on={'t':['sessionnr'], 'v':['sessionnr'],
'a':['sessionnr'], 'dc':['sessionnr'], 'z':['sessionnr']},
p_outlier=.05)
# ============================================ #
# do the actual sampling
# ============================================ #
model_filename = os.path.join(thispath, 'modelfit-md%d'%trace_id)
m.sample(nsamples, burn=nsamples/4, thin=3, db='pickle',
dbname=model_filename)
m.save(model_filename) # save the model, see if this works with pickle
# ============================================ #
# save the output values
# ============================================ #
results = m.gen_stats()
results.to_csv(os.path.join(thispath, 'results-md%d.csv'%trace_id))
# save the DIC for this model
text_file = open(os.path.join(thispath, 'DIC-md%d.txt'%trace_id), 'w')
text_file.write("Model {}: {}\n".format(trace_id, m.dic))
text_file.close()
# plot some output stuff in figures subfolder
figpath = os.path.join(thispath, 'figures-md%d'%trace_id)
if not os.path.exists(figpath):
os.mkdir(figpath)
m.plot_posteriors(save=True, path=figpath, format='pdf')
import os
import hddm
from patsy import dmatrix # for generation of (regression) design matrices
import numpy as np # for basic matrix operations
model_dir = os.getcwd()
############################################
### 1. FIT THE MODEL #######################
############################################
data = hddm.load_csv("EEGdata_hddm.csv")
samples = 10000 #run 10 models of 1000O
data.prevpe_bin_noern[data.prevpe_bin_noern==1] = 'A_bin' #this becomes the reference
data.prevpe_bin_noern[data.prevpe_bin_noern==2] = 'B_bin' #
data.prevpe_bin_noern[data.prevpe_bin_noern==3] = 'C_bin' #
data.prevpe_bin_noern[data.prevpe_bin_noern==4] = 'D_bin' #
data.prevpe_bin_noern[data.prevpe_bin_noern==5] = 'E_bin' #
data.postpe_bin_noern[data.postpe_bin_noern==1] = 'A_bin' #this becomes the reference
data.postpe_bin_noern[data.postpe_bin_noern==2] = 'B_bin' #
data.postpe_bin_noern[data.postpe_bin_noern==3] = 'C_bin' #
data.postpe_bin_noern[data.postpe_bin_noern==4] = 'D_bin' #
data.postpe_bin_noern[data.postpe_bin_noern==5] = 'E_bin' #
#define the specific link functions for z and v
def z_link_func(x, data=data):
stim = (np.asarray(dmatrix('0 + C(s,[[1],[-1]])', {'s':data.stimulus.ix[x.index]})))
return 1 / (1 + np.exp(-(x * stim)))
