Knime server: collecting feedback for supervised machine learning


I would need to know if it is possible to use knime web portal to display the results of a machine learning classification algorithm to the end user and to allow the user with a drop down menu to give a feedback selecting if the classification is correct or not for each case/record/prediciton.

The feedback from the end user should be stored in Knime and used to improve the algorithm.

Each page of the web interface should show the data of one case/record/prediction in a nice interface that allow the user to navigate to the following case/record/prediction to check it.

And when the feedback of the user is collected, next time he opens the web interface the prediction already checked by the user should not appear anymore.

Thanks in advance for this information


P.S. I would need an interface to allow the end user to provide feedbacks, whenever he wants and not necessary for all the case/record/prediction at the same time, so Active Learn nodes should not be not an option.


Hi iiiaaa,

This sounds like a very interesting use case. Do you have a rough estimate yet of the potential scale of this? E.g., do you plan to double-check each prediction, or just a sample? In general, your idea can be implemented using the KNIME WebPortal. One way to approach this could be to store the prediction results in a database and then show a number of these predictions to the user each time the workflow is run. The records shown in each run could then be updated in the database so that they are not shown the next time.

Interactivity in the form of e.g. text input can be placed on each page so your users can provide feedback if this is desired.

By the way, did you know that there is now a public KNIME WebPortal? You can find it here: You can log in with your KNIME Forum account, no other registration needed. There is a selection of workflows on there that show the basic WebPortal functionality. I hope this gives you some inspiration for how your own project can be carried out on the KNIME WebPortal :-)




Very very good idea indeed ! But I can't connect...

Error message :

Bad Gateway

The proxy server received an invalid response from an upstream server.

Apache/2.4.18 (Ubuntu) Server at Port 443

Best regards

Hi Roland,

thanks for your answer. Every prediction should be checked (we are talking about 500.000 of records).

I tried to login using your link but I received this message:

Login failed. User is not in the list of allowed login groups.