Freethought & Rationalism Archive

NialScorva · 07-07-2003, 07:49 AM

Why isn't Dembski trying something like this?

Emma Peel · 07-07-2003, 04:25 PM

Well, from the way the article explains how the software works...

Quote:

At the heart of the software is a "knowledgebase" containing a huge taxonomy of the possible ways in which death can be caused and the different types of evidence and facts that relate to them - such as eyewitness accounts, medical analyses and forensic evidence taken from the scene of crime.

An investigator simply keys in pieces of evidence and the software tries to work out how the strands of evidence could be linked and formulates possible scenarios. It also calculates how likely each scenario is, using a scale of probabilities.

For example, if medical evidence points to the cause of death as being a subdural haemorrhage the software will consider all the possible causes for subdural haemorrhage in its knowledgebase.

"It could be accidental or alcohol induced," says Keppen. "In the elderly and with infants, falling over could be the cause - or it could be caused by a blow to the head." These scenarios can all be explored by examining the other evidence.

...it sounds like it relies on intimate knowledge of its subject matter to come up with its inferences about causality. The Explanatory Filter isn't supposed to need no stinkin' intimate knowledge, is it?

The fact that nobody's used the Explanatory Filter as the basis for a forensic program is either an untapped product (and software patent) opportunity, or else it's more circumstantial evidence that his Explanatory Filter is fundamentally flawed.

07-07-2003, 07:49 AM	#1
NialScorva Veteran Member Join Date: Jan 2001 Location: Median strip of DC beltway Posts: 1,888	At least someone is persuing ID seriously Why isn't Dembski trying something like this?

Thread Tools	Search this Thread
Show Printable Version	Search this Thread: Advanced Search

Freethought & Rationalism Archive

The archives are read only.