Reasoning with uncertain data [7]

Here is an expert system user interface example that elicits both the value of an attribute and the user's confidence in that value:
The smell of gasoline is:
present when trying the starter
not present when trying the starter
I don't know/would rather not answer

How confident do you feel about your response?
Very uncertain (50%) Very certain (100%)
When the inference engine evaluates a rule with uncertain premise attribute values the following methods might be considered for combining two confidence factors (CFs) in the range 0% to 100% and designated CF1 and CF2:

MethodDefinition
MinimumUse smaller of CF1, CF2
MaximumUse larger of CF1, CF2
Average(CF1 + CF2)/2
Probability SumCF1 + CF2/100 x (100-CF1)
Multiplication(CF1 x CF2)

Some examples show how the choice of method is driven by the circumstances in which CFs must be combined.


[Next] [Previous] [1 2 3 4 5 6 7 8 9 10 [Home]
Inference Methods and Uncertainty Copyright © 2001 by eXpertise2Go.com. All rights reserved.