The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
"[SOLVED] graph Visualization of model"
venkatesh20
Member Posts: 23 
Maven
Hi,
I am developing optimization based learners as a plugin for Rapid miner. In the learning , in each iteration a cost is minimized and it should decrease for every iteration if the algorithms is correct. My need is there a way to visualize a running graph with no of iterations and cost as axes while the learner is running, if this is not possible i would like to display the same graph when the model is generated. Please advise which is the way to go.
Venki
I am developing optimization based learners as a plugin for Rapid miner. In the learning , in each iteration a cost is minimized and it should decrease for every iteration if the algorithms is correct. My need is there a way to visualize a running graph with no of iterations and cost as axes while the learner is running, if this is not possible i would like to display the same graph when the model is generated. Please advise which is the way to go.
Venki
Tagged:
0
Answers
showing graphical output during model generation is against the policy of RapidMiner - if each operator could open windows the screen is easily cluttered, and what should happen if the process is executed on a headless machine?
You can however store data for the individual iterations in your model and render them in a custom model renderer. For that, you have to register a renderer for your model in ioobjects.xml. The renderer class must implement the Renderer interface; you probably want to extend AbstractRenderer.
Best,
Marius
Thanks for the reply. With some digging on the code core, i found how to use simpePlotterDialog to show a plot after an operator finishes. But there is a real time need here, in most of the optimization problems the best way to check whether the algorithm is converging for given parameters is to see the iteration-cost plot. If it is possible to support real time plotting, then we can stop the learner if we find that it is not converging for the given parameters. This will solve lot of time in practical scenarios than to wait for the learner to finish and see how it performed.
Best,
Marius