XT,coord2,count = sep(Tdata)
coord2 = scaler.fit_transform(coord2)
coord2 = np.round(coord2.dot(2))/2
new_coord2, test_coord_values, digits = turn_y_into_binary(coord2)
### Scale testing data
XT = scaler.fit_transform(XT)
### Predict
Y_pred = clf.predict(XT)
Y_pred_normalized = turn_binary_into_y(Y_pred, test_coord_values, digits)
#Y_pred = Y_pred.reshape(-1,1)
## Calculate the statistics
#e,m,s,r = calcStats(new_coord2, Y_pred,isPrint=True)
e,m,s,r = calcStats(coord2, Y_pred_normalized,isPrint=True)
line_to_write = (test_ind - 1) * len(training_trials) + train_ind + 2
print('line to write: ', line_to_write)
if line_to_write == (test_ind - 1) * len(training_trials) + 3:
worksheet.write('A' + str(line_to_write), 'test' + str(test_ind))
worksheet.write('B' + str(line_to_write), B_entry)
worksheet.write('C' + str(line_to_write), C_entry)
worksheet.write('D' + str(line_to_write), str(r))
worksheet.write('E' + str(line_to_write), str(m))
worksheet.write('F' + str(line_to_write), str(s))
elapsed = time.time() - time_training_start
print('Total Time for ' + ftrain + ' : ', elapsed,' sec')
workbook.close()
t_end = time.time() - t
### Scale testing data
XTr = XT[:, Xi]
YTr = XT[:, Yi]
XTr = scalerX.fit_transform(XTr)
YTr = scalerY.fit_transform(YTr)
### Predict
Y_pred1 = clfX.predict(XTr)
Y_pred2 = clfY.predict(YTr)
# coord2_pred = np.concatenate((Y_pred1, Y_pred2), axis=0)
coord2_pred = np.vstack((Y_pred1, Y_pred2))
coord2_pred = coord2_pred.T
## Calculate the statistics
#e,m,s,r = calcStats(new_coord2, Y_pred,isPrint=True)
e, m, s, r = calcStats(coord2, coord2_pred, isPrint=False)
line_to_write = (test_ind -
1) * len(training_trials) + train_ind + 2
print('line to write: ', line_to_write)
print(' ')
if line_to_write == (test_ind - 1) * len(training_trials) + 3:
worksheet.write('A' + str(line_to_write),
'test' + str(test_ind))
worksheet.write('B' + str(line_to_write), B_entry)
worksheet.write('C' + str(line_to_write), C_entry)
worksheet.write('D' + str(line_to_write), str(r))
worksheet.write('G' + str(line_to_write), str(m))
worksheet.write('H' + str(line_to_write), str(s))
e, m, s, r = calcStatsOneAxis(coord2[:, 0],
Y_pred1,
Y_pred = clf.predict(XT)
# if digits_test != digits_train:
# print('Digits Mismatch')
# Y_pred_original = turn_binary_into_y_new(Y_pred, train_coord_values, digits_train)
## Y_pred = extend_y_pred(Y_pred, train_coord_values, digits_train, test_coord_values, digits_test)
## Y_pred_normalized = turn_binary_into_y(Y_pred, test_coord_values, digits)
# elif digits_test == digits_train:
# print('Digits Match')
coord2_pred = turn_binary_into_y_new(Y_pred, test_coord_values,
digits_test)
#Y_pred = Y_pred.reshape(-1,1)
## Calculate the statistics
#e,m,s,r = calcStats(new_coord2, Y_pred,isPrint=True)
e, m, s, r = calcStats(coord2, coord2_pred, isPrint=True)
line_to_write = (test_ind -
1) * len(training_trials) + train_ind + 2
print('line to write: ', line_to_write)
print(' ')
if line_to_write == (test_ind - 1) * len(training_trials) + 3:
worksheet.write('A' + str(line_to_write),
'test' + str(test_ind))
worksheet.write('B' + str(line_to_write), B_entry)
worksheet.write('C' + str(line_to_write), C_entry)
worksheet.write('D' + str(line_to_write), str(r))
worksheet.write('E' + str(line_to_write), str(m))
worksheet.write('F' + str(line_to_write), str(s))
elapsed = time.time() - time_training_start
coord2 = scaler3.fit_transform(coord2)
coord2 = np.round(coord2.dot(2)) / 2
new_coord2, test_coord_values, digits = turn_y_into_binary(
coord2)
### Scale testing data
XT = scaler.fit_transform(XT)
### Predict
Y_pred = clf.predict(XT)
Y_pred_normalized = turn_binary_into_y(Y_pred,
test_coord_values,
digits)
Y_inverse_norm = scaler3.inverse_transform(Y_pred_normalized)
e, m, s, r = calcStats(original_coord2,
Y_inverse_norm,
isPrint=True)
# Y_pred_original = turn_binary_into_y(Y_pred, original_coord2, digits)
#Y_pred = Y_pred.reshape(-1,1)
## Calculate the statistics
#e,m,s,r = calcStats(new_coord2, Y_pred,isPrint=True)
# e,m,s,r = calcStats(coord2, Y_pred_normalized,isPrint=True)
line_to_write = (test_ind -
1) * len(training_trials) + train_ind + 2
print('line to write: ', line_to_write)
print(' ')
if line_to_write == (test_ind - 1) * len(training_trials) + 3:
worksheet.write('A' + str(line_to_write),
'test' + str(test_ind))
worksheet.write('B' + str(line_to_write), B_entry)
## Scale testing data
XTr = scalerX.transform(XTr)
YTr = scalerY.transform(YTr)
## Predict
CoordX_pred = lassoX.predict(XTr)
CoordY_pred = lassoY.predict(YTr)
print("Test1: ")
## Calculate the statistics
e, m, s, r = calcStatsOneAxis(coordT[:, 0], CoordX_pred, isPrint=True)
e, m, s, r = calcStatsOneAxis(coordT[:, 1], CoordY_pred, isPrint=True)
CoordXY_pred = np.vstack((CoordX_pred, CoordY_pred)).T
e, m, s, r = calcStats(coordT, CoordXY_pred, isPrint=True)
ftest = "test2.npy"
f = 'data/robot/test/' + os.path.altsep + ftest
Tdata = np.load(f) #nxd
XT, coordT, count = sep(Tdata)
XTr = XT[:, Xi]
YTr = XT[:, Yi]
## Get cubic fit
#ply = PolynomialFeatures(3)
XTr = plyX.transform(XTr) # nxd
YTr = plyY.transform(YTr) # nxd
## Scale testing data
XTr = scalerX.transform(XTr)