beta_Val[j, 0], beta_Val[j][1] + ans[0], beta_Val[j][2] + ans[1],
beta_Val[j][3] + ans[2], beta_Val[j + 1][4]
]
sim_net2 = np.append(sim_net2, [ansK], axis=0)
sim_net2 = np.append(sim_net2, [beta_Val[12]], axis=0)
for j in xrange(12, 23):
ans = network_2.activate((beta_Val[j]))
ansK = [
beta_Val[j, 0], beta_Val[j][1] + ans[0], beta_Val[j][2] + ans[1],
beta_Val[j][3] + ans[2], beta_Val[j + 1][4]
]
sim_net2 = np.append(sim_net2, [ansK], axis=0)
#denormalize and rescale to real size
sim_deNorm2 = np.column_stack((de_normalizer(sim_net2[:, 0], theta_Val[:, 0]),
de_normalizer(sim_net2[:, 1], theta_Val[:, 1]),
de_normalizer(sim_net2[:, 2], theta_Val[:, 2]),
de_normalizer(sim_net2[:, 3], theta_Val[:, 3]),
de_normalizer(sim_net2[:, 4], theta_Val[:, 4])))
temp2 = pd.DataFrame(sim_net2, columns=lista_col3)
temp2a = pd.DataFrame(sim_deNorm2, columns=lista_col3)
beta = pd.concat([beta, temp2], axis=1, keys=['Experimental', 'Simulated'])
theta = pd.concat([theta, temp2a], axis=1, keys=['Experimental', 'Simulated'])
#save
beta.to_csv(save_fold + 'Lutein_NN_Type2H1_CrossValNew_LessNoise2_ExpTraj.csv')
theta.to_csv(save_fold +
'Lutein_NN_Type2H1_CrossValNew_LessNoise2_ExpTraj_Rescale.csv')
#overwrite
]
sim_net2 = np.append(sim_net2, [ansK], axis=0)
sim_net2 = np.append(sim_net2, [beta_Val[12]], axis=0)
for j in xrange(12, 23):
ans = network_2.activate((beta_Val[j]))
ansK = [
beta_Val[j, 0], beta_Val[j][1] + ans[0], beta_Val[j][2] + ans[1],
beta_Val[j][3] + ans[2], beta_Val[j + 1][4]
]
sim_net2 = np.append(sim_net2, [ansK], axis=0)
#denormalize and rescale to real size
sim_deNorm2 = np.column_stack(
(de_standardizer(
de_normalizer(sim_net2[:, 0], standardizer(theta_Val[:, 0])),
theta_Val[:, 0]),
de_standardizer(
de_normalizer(sim_net2[:, 1], standardizer(theta_Val[:, 1])),
theta_Val[:, 1]),
de_standardizer(
de_normalizer(sim_net2[:, 2], standardizer(theta_Val[:, 2])),
theta_Val[:, 2]),
de_standardizer(
de_normalizer(sim_net2[:, 3], standardizer(theta_Val[:, 3])),
theta_Val[:, 3]),
de_standardizer(
de_normalizer(sim_net2[:, 4], standardizer(theta_Val[:, 4])),
theta_Val[:, 4])))
temp2 = pd.DataFrame(sim_net2, columns=lista_col3)
#denormalize and rescale to real size
#print 'sim ', sim_net2[0,:]
#print ' iota', iota[0,:]
#print 'theta', theta_Val[0,:]
#print 'arr temp', arr_temp[0,:]
#print de_normalizer(sim_net2[:,3],standardizer(iota[:,3]))
#print de_standardizer(de_normalizer(sim_net2[:,3],standardizer(iota[:,3])),iota[:,3])
#print iota[:,3]
#print standardizer(iota[:,3])
#print normalizer(standardizer(iota[:,3]))
#sys.exit()
sim_deNorm2 = np.column_stack(
(sim_net2[:, 0],
de_standardizer(de_normalizer(sim_net2[:, 1], standardizer(iota[:, 1])),
iota[:, 1]),
de_standardizer(de_normalizer(sim_net2[:, 2], standardizer(iota[:, 2])),
iota[:, 2]),
de_standardizer(de_normalizer(sim_net2[:, 3], standardizer(iota[:, 3])),
iota[:, 3]),
de_standardizer(de_normalizer(sim_net2[:, 4], standardizer(iota[:, 4])),
iota[:, 4])))
#de_standardizer(de_normalizer(arr_tempA[:,2],standardizer(iota[:,2])),iota[:,2])
temp2 = pd.DataFrame(sim_net2, columns=lista_col3)
temp2a = pd.DataFrame(sim_deNorm2, columns=lista_col3)
beta = pd.concat([beta, temp2], axis=1, keys=['Experimental', 'Simulated'])
theta = pd.concat([theta, temp2a], axis=1, keys=['Experimental', 'Simulated'])
for j in xrange(12, 23):
ans = network_2.activate((beta_Val[j]))
ansK = [
beta_Val[j, 0], beta_Val[j][1] + ans[0], beta_Val[j][2] + ans[1],
beta_Val[j][3] + ans[2], beta_Val[j + 1][4]
]
sim_net2 = np.append(sim_net2, [ansK], axis=0)
#denormalize and rescale to real size
#print theta_Val
#print beta_Val
#sys.exit()
dataSet1 = np.column_stack(
(sim_net2[:sLht, 0],
de_standardizer(
de_normalizer(sim_net2[:sLht, 1], standardizer(arr_temp[:sLht, 1])),
arr_temp[:sLht, 1]),
de_standardizer(
de_normalizer(sim_net2[:sLht, 2], standardizer(arr_temp[:sLht, 2])),
arr_temp[:sLht, 2]),
de_standardizer(
de_normalizer(sim_net2[:sLht, 3], standardizer(arr_temp[:sLht, 3])),
arr_temp[:sLht, 3]),
de_standardizer(
de_normalizer(sim_net2[:sLht, 4], standardizer(arr_temp[:sLht, 4])),
arr_temp[:sLht, 4])))
dataSet2 = np.column_stack(
(sim_net2[sLht:, 0],
de_standardizer(
de_normalizer(sim_net2[sLht:, 1], standardizer(arr_temp[sLht:, 1])),
#simulation for Network 2, 1 Hidden.
for j in xrange(11):
ans=network_2.activate((beta_Val[j]))
ansK=[beta_Val[j,0],beta_Val[j][1]+ans[0],beta_Val[j][2]+ans[1],beta_Val[j][3]+ans[2],beta_Val[j+1][4]]
sim_net2=np.append(sim_net2,[ansK],axis=0)
sim_net2=np.append(sim_net2,[beta_Val[12]],axis=0)
for j in xrange(12,23):
ans=network_2.activate((beta_Val[j]))
ansK=[beta_Val[j,0],beta_Val[j][1]+ans[0],beta_Val[j][2]+ans[1],beta_Val[j][3]+ans[2],beta_Val[j+1][4]]
sim_net2=np.append(sim_net2,[ansK],axis=0)
#denormalize and rescale to real size
print sim_net2[:,3]
print standardizer(theta_Val[:,3])
print de_normalizer(sim_net2[:,3],standardizer(theta_Val[:,3]))
print de_standardizer(de_normalizer(sim_net2[:,3],standardizer(theta_Val[:,3])),theta_Val[:,3])
print theta_Val[:,3]
#sys.exit()
sim_deNorm2=np.column_stack((
sim_net2[:,0]
,de_standardizer(de_normalizer(sim_net2[:,1],standardizer(theta_Val[:,1])),theta_Val[:,1])
,de_standardizer(de_normalizer(sim_net2[:,2],standardizer(theta_Val[:,2])),theta_Val[:,2])
,de_standardizer(de_normalizer(sim_net2[:,3],standardizer(theta_Val[:,3])),theta_Val[:,3])
,de_standardizer(de_normalizer(sim_net2[:,4],standardizer(theta_Val[:,4])),theta_Val[:,4])))
#de_standardizer(de_normalizer(arr_tempA[:,2],standardizer(arr_temp[:,2])),arr_temp[:,2])
temp2=pd.DataFrame(sim_net2,columns=lista_col3)
temp2a=pd.DataFrame(sim_deNorm2,columns=lista_col3)