Converting Datasets to Structures

The Convert Dataset to Structure task can produces a number of structures from a dataset in input. Those structures include replacement rules, autoregressive models, cluster labels, clusters, discretization cutoffs, frequent itemsets, frequent sequences, pca eigenvectors, results, rules, models, and pca eigenvectors.


There are many reasons why you may want to convert structures, here are a few examples:

  • to quickly add a large number of heuristic rules into a process, by  inserting the rules into a table, which can then be imported into a process as a dataset, and then converted into a ruleset.

  • to reconvert structures that were previously converted into a dataset, using the Convert Structure to Dataset task, in order to perform in-depth analysis in the Data Manager, back into their original format.

  • to create a model from a dataset, which can then be used in an Apply Model task to derive its responses in correspondence of given samples.


The more specific Convert Dataset to Ruleset (and vice-versa) and Convert Dataset to Model (and vice-versa) tasks are still available for backward compatibility, but they are fully substituted by this new set of generic structure conversion tasks.


Prerequisites


Procedure

  1. Drag and drop the Convert Dataset to Structure task onto the stage.

  2. Connect a task that contains an existing dataset to the Convert Dataset to Structure task.

  3. Double click the task and select the required structure. The only structure that requires additional parameters are rules, which require the parameters explained in the table below.

  4. Save and compute the task. 

Convert Dataset to Ruleset options

Parameter Name

PO

Description

Rule ID atttibute

ruleidname

Select the attribute in the dataset which contains the ID for each rule. These numbers must be unique and consecutive. If this is not the case, leave this option blank and new numbers will be assigned to the rows. 

Rule output name attribute

ruleoutatname

Select the attribute in the dataset that contains the output attributes.

Rule output value attribute

ruleoutvalname

Select the attribute in the dataset that contains the output attribute values.

Rule covering attribute

rulecovname

Select the attribute in the dataset that contains the covering value.

Rule error attribute

ruleerrname

Select the attribute in the dataset that contains the error value.

Rule conditions (NOMINAL)

condnames

Drag and drop here all the attributes that represent rule conditions.
Instead of manually dragging and dropping attributes, they can be defined via a filtered list. This is preferable if you have many attributes, and you select, for example, all attributes that start with the word Condition.