Components > Refining intelligent structure models

Refining intelligent structure models

After Intelligent Structure Discovery discovers the structure of the model input, you can refine the structure so that when you use the model in production the output meets your requirements.
For example, rename a node, add a prefix to file names, or define rows or columns in Microsoft Excel files as table headers.
Use the Visual Model and Relational Output tabs of the Intelligent Structure Model page to understand and refine the output.
On the Visual Model tab, you can see the output in a tree-like structure. The model shows the discovered types of data as nodes, and shows their relationship to each other in a graphical format. Use the Visual Model tab to trace how the input data is mapped to a node.
On the Relational Output tab, you can see the relational output that the intelligent structure produces. The output is organized in one or more output groups. An output group contains one or more nodes. Use the Relational Output tab to determine how a node is mapped to an output group.
When you refine the discovered structure, you can test the output that Intelligent Structure Discovery generates based on the model in different output formats. Repeat refining and testing the model until you are satisfied that it generates the required output.
The following image shows the process of discovering, refining, testing, and saving a model:
The image shows four elements that represent the stages of the process and the relationships between them. The first stage is to discover the structure, which leads to the second stage- testing the output. This stage leads to a decision point: is the output OK? If yes, the decision point leads directly to the last stage- saving the model. If not, the decision point leads to the third stage- refining the model. The process of testing the output and refining the model repeats until the output is OK, and then the decision point leads to the fourth and last stage- saving the model.