|
|
| Causal Analytics |
|
Types of Causal
Relationships |
 |
A truly drives B. Every time that A moves, so will B. This is a true cause and effect relationship. |
 |
A drives both B and C, but A is hidden (not in the data), so B appears to cause C. This is a false causal relationship that can be a challenge to discern from pure data analytics. It can be seen only if B is independently moved either in the data or through a test. In those cases, it will be seen that C does not actually respond to B. |
 |
A drives B and D drives C. If both A and D are hidden, there may be times that B appears to influence C. However, this is easier to deduce because the B-to-C relationship will be infrequent and/or inconsistent. |
 |
A drives B and in turn B drives A. This type of feedback loop is seen frequently. For example, an operator changes the temperature of a manufacturing process (A), which changes the product characteristics (B). The operator sees these characteristics and then decides to change the temperature again (A). |
|
These are simple examples.
In typical applications, different causal relationships are combined into
more complex systems. The challenge is to identify and quantify
these relations to use them to our advantage. There are a number of
useful analytical methods to do this. |
For more information, please contact us at tech@biocompsystems.com
Legal
Policy | Privacy
Policy | Contact Us
Copyright © 1995-2007 BioComp Systems, Inc.. All rights reserved.
IntelliDynamics is a registered trademark of BioComp Systems, Inc.
Last modified:
Sunday April 24, 2005.
Serving the web since Feb 2, 1996