Importing Structures with Import from Task

You can import many different types of structure, such as models, clusters and rules, from any part of your current process, or from the tasks belonging to any other process, as long as they use the same underlying Rulex database.

This is useful to help create simple processes, minimizing the connections between various tasks, and also to share information between different processes.

There are some important advantages to working in this way:

You can create different pre-processing tasks in the new process
You can join or relate different training models in a single model
You can construct loops or multiple iterations.

We do not recommend using this task within modules. As data sources change, if unexpected results are produced as output, it becomes very difficult to retrieve the root of the issue.

Prerequisites

You must have created and activated a process in Rulex.
The processes you import data from must all use the same underlying database.

Additional tabs

The following additional tabs are provided:

Documentation tab where you can document your task,
Parametric options tab where you can configure process variables instead of fixed values. Parametric equivalents are expressed in italics in this page (PO).

Procedure

Drag and drop the Import from Task task onto the central stage.
Double click the task.
Configure the task options, as described in the table below.
Save and compute the task.

Parameter Name	PO	Description
Import from Task options
Import dataset from	setref	Select the type of item from which you want to create your new dataset. These are usually based on other datasets, however, in some cases it is useful to use items such as clusters or frequent itemsets as the basis of the new dataset.
Structures to be imported from target task	reflist	Select which structures you want to include in the imported dataset. Possible structures, along with the tasks that produce these structures are:
		Association rules	Hierarchical Basket Analysis Similar Items Detector
		Auto regressive models	Auto-Regressive
		Clusters	Label Clustering Projection Clustering Standard Clustering
		Cluster labels	Label Clustering Projection Clustering
		Discretization cutoffs	Discretize
		Frequent itemsets	Frequent Itemsets Mining
		Frequent sequences	Sequences Analysis
		Results	All tasks
		Rules	LLM tasks (Classification, Regression & One-Class) Decision Tree Regression Tree
		Models	Logistic Linear Neural Networks (Classification & Regression) SVM (Classification & Regression)
		PCA eigenvectors	Principal Component Analysis
Process	process	Select the specific process from which you want to import the structures from the drop-down list. If ‘THIS’ is selected within modules, datasets coming from the parent workflow are imported. However, it is preferable to map inputs in the module itself, rather than using Import from Task, to avoid possible inconsistencies.
Task	task	Select the specific task you want to import data from.