Training quest
The process includes machine learning phases that can have multiple steps running in a constructed sequence. You can train multiple models in parallel inside one quest. Up to four models can be compared by viewing the results in one screen to decide which model works best for your needs.
Each individual activity is a computational unit in the flow that has an output result. To see the output results, you can click the output port of the activity box. You can track the metrics of each step and adjust accordingly.
Use distinct steps to fine-tune the training flow. Then you can rerun only the steps you require during the testing phase. Data sources and intermediate data are reused across the flow.
Use the catalog of pre-built algorithms and modules to create and manage multiple simple and complex flows.