The "Compute Clustering" tab allows the user to select the clustering that will be applied on the dataset.
Figure 6: Pre compute tab empty.
Figure 7: Pre compute tab with some clustering method selected.
The button panel contains 3 buttons:
- The first button "Start" allows to launch the clustering that have been selected and moves on to the next step which is the result when all the clustering are done.
- The second button "Select All" allows to select all the clustering methods with a single click. Warning the more clustering methods have been selected, the longer it will take. If the application is hosted on a server, it will depend on the server capacity. Otherwise, if it is local, it will depend on the capacity of the computer on which you run the application.
- The third button "Unselect All" allows to deselect all the methods that have been selected.
The panel on the right serves as a panel to display the progress of the clustering. At each step, a statement about the progress will be displayed.
Number of cluster to compute panel
This panel is just to inform the number of clusters to cut.If the user doesn't know in advance the number of clusters to obtain, he can leave the value on 0. When the value is 0, the PEV method is applied to find the number of clusters to compute. This is only necessary for algorithms that require the K value of the number of clusters as input.
Selection of the clustering methods
To choose a classification method to apply, simply check the checkbox associated with the method. For example, to perform Spectral PAM clustering, you just have to check the checkbox next to "Compute Spectral PAM Clustering".
When a checkbox has been checked, if there are additional options related to the clustering method, these will appear and the user can then modify them as needed.