import sys
import os
import Globals
import numpy as np
from matplotlib import pyplot as plt
OUTPUT_PATH = os.path.expanduser("~") + "/" + "Python_outputs"
if not os.path.exists(OUTPUT_PATH):
os.makedirs(OUTPUT_PATH)
OUTPUT_PATH = "/home/ubuntu-lieder-pc/Python_outputs/"
OUTPUT_PATH = "/home/dell/Python_outputs/"
CODE_PATH = Globals.path_dic("code")
DATA_PATH = Globals.path_dic("data")
# Make sure the data files are named: 'continuous.mat'
# and 'bimodal.mat' for each respective experiment
sys.path.append(CODE_PATH)
from data_loader.load_data import *
from inference_toolboxes.pymc3_functions.inference_pymc3 import *
from stimulus_generation.generate_stimuli import *
from model_simulations.descriptive_model import *
def plot_distribution(s1, s2):
f, ax = plt.subplots(2, 2)
ax[0, 0].plot(s1, s2, ".")
ax[0, 1].plot(s1[1:] - s2[1:], s1[1:] - 0.5 * (s1[:-1] + s2[:-1]), ".")
ax[1, 0].hist(s1, bins=30)
ax[1, 1].hist(s2, bins=30)
for i_typ, typ in enumerate(["good", "poor"]):
save_prefix = "linear_part/"
if not os.path.exists("./" + save_prefix):
os.makedirs("./" + save_prefix)
save_name = model + "_ml_fit_" + typ + "_" + str(thresh) + ".svg"
print save_name
# =========================================
# Loading data per accuracy
# =========================================
data_path = Globals.path_dic("data")
loader = Dataloader(data_path + "continuous.mat")
poor = (0.65, 0.85)
good = (0.85, 1)
r = poor if typ == "poor" else good
I = loader.subject_indices_for_acc_range(r)
F1, F2, Y = loader.subject_data_from_indices(list(I))
# =========================================
# Preprocess data for regression
# =========================================
x, y = get_trial_covariates(F1, F2, Y, T=T, inf=0)
x = x.T
import sys
import os
import Globals
from matplotlib import pyplot as plt
OUTPUT_PATH = os.path.expanduser('~')+'/'+'Python_outputs'
if not os.path.exists(OUTPUT_PATH):
os.makedirs(OUTPUT_PATH)
OUTPUT_PATH = "/home/ubuntu-lieder-pc/Python_outputs/"
CODE_PATH = Globals.path_dic('code')
DATA_PATH = Globals.path_dic('data')
CODE_PATH = "/home/ubuntu-lieder-pc/git/additive_modelling_pitch_bias/"
DATA_PATH = CODE_PATH + "data_files/"
print CODE_PATH
# Make sure the data files are named: 'continuous.mat'
# and 'bimodal.mat' for each respective experiment
sys.path.append(CODE_PATH)
from data_loader.load_data import *
from inference_toolboxes.pymc3_functions.inference_pymc3 import *
samples=10000
T = 1
inf = 0
dataloader = Dataloader(DATA_PATH +'continuous.mat')
groups = 3
acc_sample = np.array([[0.63,0.75],[0.75,0.88],[0.88,1]])
# iterating over lags
for T in [1,2]:
# iterating over groups of subjects
for i_typ,typ in enumerate(['good','poor']):
save_name = mod_typ+'_'+typ+str(T)+'_tresh_'+str(thresh)+'.svg'
print save_name
#=========================================
# Loading data per accuracy
#=========================================
data_path = Globals.path_dic('data')
loader = Dataloader(data_path+'continuous.mat')
r = poor if typ == 'poor' else good
I = loader.subject_indices_for_acc_range(r)
F1,F2,Y = loader.subject_data_from_indices(list(I))
#=========================================
# Preprocess data for regression
#=========================================
x,y = get_trial_covariates(F1,F2,Y,T=T,inf=0)
x = x.T
#=========================================
# Subselect close trials
