Azure机器学习餐厅推荐器Web服务不起作用

Question

I would like to build a recommender very similar to the Azure ML restaurant recommender sample. However if i use the sample as-is, I always get 'empty output' returned from the webservice. These are steps i followed.

1. Create the training experiment from the sample and run it. Results look good.
2. Create the predictive experiment and run it. Again visualising data looks good.
3. Publish the web service.
4. Enter test useid,placeid,rating with various values picked from the test data. Always return 'empty output'. I made no changes to the experiments along the way, so this is what the predictive looks like:

What am I doing wrong?

Answer 1

Got it working. I figured out the webservice input should be connected to the customer input, as i am making recommendations for a customer. Makes sense.