fix: add correct ex vars to cross validation

[AzulGarza]

This PR adds exogenous variables correctly to the cross_validation method.
Before the change, the exogenous variables feature was tested using the argument date_features.
However, an actual exogenous variable wasn't tested.
This PR fixes that.
Also, now the SDK prints the exogenous variables sent to the API.

Note: the bug from the initial inclusion of the exogenous variables
came from the confusion between df and X_df for the SDK and the API.
See this explanation.
Currently, the code does not reflect this difference.

More on this: the models_performance test didn't reflect this problem because currently, it does not use cross_validation. O: See #160.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[github-actions]

Experiment 1: air-passengers

Description:

variable	experiment
h	12
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	12.6793	11.0623	47.8333	76
mape	0.027	0.0232	0.0999	0.1425
mse	213.936	199.132	2571.33	10604.2
total_time	1.796	2.1218	8.7664	0.7792

Plot:

Experiment 2: air-passengers

Description:

variable	experiment
h	24
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	58.1031	58.4587	71.25	115.25
mape	0.1257	0.1267	0.1552	0.2358
mse	4040.21	4110.79	5928.17	18859.2
total_time	2.1684	2.6309	0.0078	0.0075

Plot:

Experiment 3: electricity-multiple-series

Description:

variable	experiment
h	24
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	142.394	196.363	269.23	1331.02
mape	0.0203	0.0234	0.0304	0.1692
mse	63464.7	123119	213677	4.68961e+06
total_time	3.1307	2.7136	0.0189	0.019

Plot:

Experiment 4: electricity-multiple-series

Description:

variable	experiment
h	168
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	522.427	353.528	398.956	1119.26
mape	0.069	0.0454	0.0512	0.1583
mse	966294	422332	656723	3.17316e+06
total_time	3.5122	4.2689	0.0204	0.0192

Plot:

Experiment 5: electricity-multiple-series

Description:

variable	experiment
h	336
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	478.362	361.033	602.926	1340.95
mape	0.0622	0.046	0.0787	0.17
mse	805039	441118	1.61572e+06	6.04619e+06
total_time	5.023	2.9784	0.02	0.0204

Plot:

[github-actions]

Experiment 1: air-passengers

Description:

variable	experiment
h	12
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	12.6793	11.0623	47.8333	76
mape	0.027	0.0232	0.0999	0.1425
mse	213.936	199.132	2571.33	10604.2
total_time	4.0352	3.5667	7.4783	0.676

Plot:

Experiment 2: air-passengers

Description:

variable	experiment
h	24
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	58.1031	58.4587	71.25	115.25
mape	0.1257	0.1267	0.1552	0.2358
mse	4040.21	4110.79	5928.17	18859.2
total_time	2.0667	1.9551	0.0062	0.0059

Plot:

Experiment 3: electricity-multiple-series

Description:

variable	experiment
h	24
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	142.394	196.363	269.23	1331.02
mape	0.0203	0.0234	0.0304	0.1692
mse	63464.7	123119	213677	4.68961e+06
total_time	2.374	2.86	0.0182	0.0162

Plot:

Experiment 4: electricity-multiple-series

Description:

variable	experiment
h	168
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	522.427	353.528	398.956	1119.26
mape	0.069	0.0454	0.0512	0.1583
mse	966294	422332	656723	3.17316e+06
total_time	3.774	2.7661	0.0173	0.0168

Plot:

Experiment 5: electricity-multiple-series

Description:

variable	experiment
h	336
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	478.362	361.033	602.926	1340.95
mape	0.0622	0.046	0.0787	0.17
mse	805039	441118	1.61572e+06	6.04619e+06
total_time	4.6603	2.9938	0.0182	0.0174

Plot:

[github-actions]

Experiment 1: air-passengers

Description:

variable	experiment
h	12
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	12.6793	11.0623	47.8333	76
mape	0.027	0.0232	0.0999	0.1425
mse	213.936	199.132	2571.33	10604.2
total_time	3.283	4.1158	7.614	0.6924

Plot:

Experiment 2: air-passengers

Description:

variable	experiment
h	24
season_length	12
freq	MS
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	58.1031	58.4587	71.25	115.25
mape	0.1257	0.1267	0.1552	0.2358
mse	4040.21	4110.79	5928.17	18859.2
total_time	4.1878	4.2396	0.0065	0.0061

Plot:

Experiment 3: electricity-multiple-series

Description:

variable	experiment
h	24
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	142.394	196.363	269.23	1331.02
mape	0.0203	0.0234	0.0304	0.1692
mse	63464.7	123119	213677	4.68961e+06
total_time	2.6946	2.6704	0.0188	0.0194

Plot:

Experiment 4: electricity-multiple-series

Description:

variable	experiment
h	168
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	522.427	353.528	398.956	1119.26
mape	0.069	0.0454	0.0512	0.1583
mse	966294	422332	656723	3.17316e+06
total_time	3.6077	5.2898	0.0176	0.0175

Plot:

Experiment 5: electricity-multiple-series

Description:

variable	experiment
h	336
season_length	24
freq	H
level	None
n_windows	1

Results:

metric	timegpt-1	timegpt-1-long-horizon	SeasonalNaive	Naive
mae	478.362	361.033	602.926	1340.95
mape	0.0622	0.046	0.0787	0.17
mse	805039	441118	1.61572e+06	6.04619e+06
total_time	4.7584	3.1866	0.0183	0.0179

Plot: