Forecast Commandedit

The forecast-model command will return a forecast for the given time range, and the given model. For example the command below will output data points for model named pricing-model between now and the next 24 hours:

loudml -e "forecast-model --from now --to now+1d pricing-model"

And this example will output result to stdout:

forecast(pricing-model): 100%|███████████████████████| 1/1 [00:01<00:00,  1.01s/it]
timestamp           @mean_value         loudml.mean_value
1569574800.0        None                13286.228
1569576000.0        None                12169.464
1569577200.0        None                12049.305
1569578400.0        None                10466.796
...

The forecast-model command supports the following options:

-f

(date) The from date in range query, supports Date Math notation.

-t

(date) The to date in range query, supports Date Math notation.

-s

Will save output to the output bucket

-i

Read data points from the specified input bucket

-o

Write output data points to the specified output bucket

-c

Stop forecasting if this constraint is met, ie when the target metric is either exceeded or becomes less than the specific constraint

-p

Generate forecast confidence interval with +/- 100*p% range

model_name

The model name