modifying results in python numpy arrays

Question

I have a response from a kerras prediction that looks like this (y_pred):

array([[127450.63 ],
 [181983.39 ],
 [150607.72 ],
 ...,
 [460400. ],
 [ 92920.234],
 [244455.97 ]], dtype=float32)

I need to compare the results to another array that looks like this (t_pred):

[105000. 172000. 189900. ... 131000. 132000. 188000.]

How would I go about converting array 1 to look like array 2 so I can calculate its mean_square_log_error, like this?:

mean_squared_log_error(t_pred, y_pred)

Answer 1

Use ravel() or reshape(-1) or flatten():

mean_squared_log_error(t_pred, y_pred.ravel())

Or

mean_squared_log_error(t_pred, y_pred.reshape(-1))

Or

mean_squared_log_error(t_pred, y_pred.flatten())

Example:

>>> from sklearn.metrics import mean_squared_log_error
>>> y_pred = np.array([[127450.63, 181983.39,181983.39 ]]) 
>>> t_pred = [105000., 172000., 189900.]
>>> mean_squared_log_error(t_pred, y_pred.ravel())
0.01418072635060214
>>> mean_squared_log_error(t_pred, y_pred.reshape(-1))
0.01418072635060214
>>> mean_squared_log_error(t_pred, y_pred.flatten())
0.01418072635060214