In order to add an algorithm to our experiment, you must perform the following actions:
- Select the desired option from the tool bar on the left, according to the type of algorithm we want to insert:
pre-process, method, post-process or test.
- Choose the desired algorithm from the tree structure.
- Click on the draw area.
There exists some types of algorithms according to their functionality. Let us
see as they are:
- Pre-process: they are discretization algorithms, instances selection and features selection.
As their name suggests, they are used to add a pre-processing step over the initial
data set, before applying a knowledge extraction algorithm.

- Methods: they are knowledge extraction algorithms: decision tree and rule extraction in supervised
learning, neural networks, unsupervised learning, etc...

- Post-process: they are algorithms that are inserted in final steps in order to refine the results
obtained by the previous algorithms.

- Test: they are statistical algorithms to analyze the results obtained by the previous methods.

- Results: they are features to show the results obtained by the previous methods.

Now that the algorithm is inserted, you can repeat these steps in order to add more algorithms. Once all the algorithms that you wish are inserted, you must connect them. To do so, click on the
button from the left tool bar; click on the source node to start the connection and then to the target node to end it. Repeat this action until all the nodes are correctly connected. If you need more information, please go to the Help menu and select Content. On the left tree, go to Help-Experiments Design-Experiment Graph-Connections. Here, you can see how to make a connection and the restrictions that you must keep in mind.