Azure Machine Learning - web input output

Question

I am new to Azure ML and I am trying to pass a web input to my Scorer , and my R script returns a Web output.

Input Parameters : No. of predictions to make - String type Output Parameters : Forecast - Data frame

For every input I pass , I get the same output.Can you help me out?

Thanks in advance.

Solution

Web service output generally shows only one value. So by default it fetches the first row element of the frame. In order to check the entire frame , use Batch execution or excel.

Sundari, can you publish your current experiment to the gallery (in unlisted mode if you like) and share the URL here? Then we'd be able to give feedback on the design. Thanks! — mewahl
@mewahl I think it shows the first value of the data frame. Is there a way I can pass a list of values as web output? — Sundari
Hi Sundari, the basic connectivity looks right to me. The red X on the Score Model module suggests that your sample data in Enter Data Manually does not have the right columns/column types for scoring. Do the column names/types perfectly match those for the data leaving your Edit Metadata module (i.e. the data used for training)? I could check this from a gallery listing but I understand if it's too confidential to share here. (Maybe try Azure ML Studio LiveChat to 1:1 with a data scientist?) — mewahl
Hi @mewahl , yeah its confidential. I got the red X resolved. Just have some issues with list values as web service output. I will try to connect with the LiveChat. Thank you so much for the help. :) — Sundari

neerajkh neerajkh · Accepted Answer · 2017-02-09T07:19:09

I think you should first create a model within an experiment and make sure that works. Then before publishing as web service, actually create a predictive experiment by clicking "predictive web service" as shown in the picture below

Azure Machine Learning - web input output

1 Answers