Forecast APIedit

After training is completed, you can request the forecast of future feature values using the forecast API.

POST /models/pricing_model/_forecast?from=now&to=now+1h&bg=true

Accepted parameters for the _forecast API are:

from

generate the requested data from this date

to

generate the requested data until this date

bg

true to run an asynchronous job in the background, false (default) to run in the foreground

On success it will return a job identifier if bg=true, or data points if bg=false or option is missing.

* HTTP 1.0, assume close after body
< HTTP/1.0 200 OK
< Content-Type: application/json
< Content-Length: 39
< Server: Werkzeug/0.12.2 Python/3.4.5
< Date: Fri, 14 Mar 2018 10:00:00 GMT
<
"b67c10d6-3886-4285-a19c-2b908323238a"

You can then query the job result using the Job API:

GET /jobs/b67c10d6-3886-4285-a19c-2b908323238a

The result of the above get operation will contain the forecast result, according to feature names and properties defined in your model.

{
    "id": "b67c10d6-3886-4285-a19c-2b908323238a",
    "state": "done",
    "type": "forecast",
    "result": {
        "predicted": {
            "pricing_feature": [
                1.0200526714324951,
                1.0138707160949707,
                1.0268415212631226,
                1.0200526714324951,
                1.0138707160949707,
                1.0268415212631226
            ]
        },
        "timestamps": [
            1521018000.0,
            1521018600.0,
            1521019200.0,
            1521019800.0,
            1521020400.0,
            1521021000.0
        ]
    }
}