Choice Emotionally supportive networks.


38 views
Uploaded on:
Category: News / Events
Description
Guideline based Systems . on the off chance that A then B If pump disappointment then the weight is low If pump disappointment then check oil level If power disappointment then pump disappointment Uncertainty If A (with assurance x) then B (with conviction f(x))If C (with sureness x) then B (with conviction g(x)If we now get the data that A holds with conviction an and C holds with conviction c, what is the assurance of B?.
Transcripts
Slide 1

Choice Support Systems

Slide 2

Rule based Systems if A then B If pump disappointment then the weight is low If pump disappointment then check oil level If control disappointment then pump disappointment Uncertainty If A (with assurance x) then B (with conviction f(x)) If C (with sureness x) then B (with conviction g(x) If we now get the data that A holds with conviction an and C holds with conviction c, what is the sureness of B?

Slide 3

Rule based Systems – cont. On the off chance that blood glucose is low before lunch, then take less insulin in the morning Model of the specialist Easy to fabricate ? Simple to keep up ? Straightforward for clinicians and patients ? Issues with instability and inconstancy

Slide 4

Rule based Systems – cont.

Slide 5

Neural Networks

Slide 6

Bayesian Networks If tonsillitis then P(temp>37.9) = 0.75 If whooping hack then P(temp>37.9) = 0.65 One could be prompt to peruse this as guidelines. They shouldn\'t be. So an alternate documentation is utilized: P(temp>37.9 | whooping hack) = 0.65 P(temp>37.9 | whooping hack, tonsillitis)

Slide 7

Bayesian Networks - cont Bayes\' hypothesis: P(A | B)P(B) = P(B | A)P(A) utilizes a worldwide viewpoint figures the new probabilities accurately in manage based frameworks you attempt to display the specialists method for thinking (henceforth the name master frameworks), while with Bayesian systems you attempt to model conditions in the area itself

Slide 8

Example diabetes Predictions of blood glucose levels in light of numerical models of the starch digestion system Illustrate the impact of evolving e.g. insulin Model of the patient Can deal with instability and fluctuation ? Issues with different components, e.g. stretch, fever, liquor, practice and so forth

Slide 9

Comparing Neural Networks and Bayesian Networks The principal distinction between the two sorts of systems is that a perceptrone in the concealed layers does not in itself have an elucidation in the space of the framework, while every one of the hubs of a Bayesian system speak to ideas that are very much characterized as for the area.

Recommended
View more...