Developing better products faster


   


FormRules Enhancements

The latest release (January 2006) of FormRules has many new features that make it easier to use.

First, there is a new Welcome screen. No longer do you simply get a blank spreadsheet. Now, a window appears to allow you either to launch a New task, or to open an existing one. If you pick Create NewTask, you get the blank spreadsheet (as before) so you can import your data. If you choose to Open an existing task, now you will get a 'browse' window that allows you to find the task you want. See the new Welcome screen below.


FormRules Welcome Screen

As the screen shows, we've also added an easy-to-access Tutorial. If you pick View Tutorial, FormRules launches a New task and displays the Tutorial Help information.

Speaking of Help - new users might want additional Help while they are getting familiar with the program. So, we have added on-screen Tips that point out what is important about each screen. As you move from screen to screen, the Tips change so that their information relates to that particular screen. Tips are always displayed on top of the main screen, but you can 'iconize' the Tips window if you wish to remove it temporarily. Once you are familiar with FormRules, you can switch the Tips off completely if you wish. Full on-line help is available via the Help buttons as always, whether or not Tips are displayed.But this added Help is not the only change. We have added some new features that make it easier to see when you've got good models, and which inputs make the biggest contributions to models and rules. Here are some examples:

The ANOVA statistics measure R2 is now displayed on the Graphical representation of the models. See an example here. R2 is shown against a colour-coded background. If the background is green (values of 70% to 99%) then you have a good model. Values between 50% and 70% are shown on a yellow background. Values below 50% are shown in red - proceed at your own risk!! Values above 99% are also shown in red, since probably the model has over-trained unless you have very clean data.

Graphical Representation of Model

The other thing you will notice from the plot is that one of the lines is coloured in purple. This shows that this submodel makes the largest positive and largest negative (or smallest positive) contribution to the overall model. The actual rules from SubModel1 for this case are:

IF MF% is LOW THEN KnoopHard is LOW (1.00)

IF MF% is HIGH THEN KnoopHard is HIGH (1.00)

All these changes should enhance your experience of using FormRules, making it more obvious exactly what information is provided by this powerful data mining program.

 

 

This document maintained by webmaster@intelligensys.co.uk.
Copyright © 2006 Intelligensys Ltd