02.03.2013 Views

Downloadable - About University

Downloadable - About University

Downloadable - About University

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Assessing the value of new information 231<br />

this indication using the probability tree in Figure 8.10(a). The prior<br />

probability of the virus being present and the test correctly indicating<br />

this is 0.3 × 0.9, which is 0.27. Similarly, the probability of the virus being<br />

absent and the test incorrectly indicating its presence is 0.7 × 0.2, which<br />

is 0.14. This means that the total probability that the test will indicate<br />

that the virus is present is 0.27 + 0.14, which is 0.41. We can now put<br />

this probability onto the decision tree (Figure 8.11).<br />

We can also use the probability tree in Figure 8.10(a) to calculate the<br />

posterior probabilities of the virus being present or absent if the test gives<br />

an indication of its presence. Using Bayes’ theorem, it can be seen that<br />

these probabilities are 0.66 and 0.34, respectively. These probabilities<br />

can also be added to the decision tree.<br />

Prior probs Conditional probs Joint probs Posterior probs<br />

0.27 0.27<br />

= 0.66<br />

0.41<br />

0.3<br />

Virus present<br />

Virus absent<br />

0.7<br />

0.3<br />

Virus present<br />

Virus absent<br />

0.7<br />

0.9<br />

Test indicates present<br />

0.2<br />

Test indicates present<br />

p(test indicates virus present) = 0.41<br />

(a)<br />

0.14 0.14<br />

= 0.34<br />

0.41<br />

Prior probs Conditional probs Joint probs Posterior probs<br />

Test indicates absent<br />

0.1<br />

Test indicates absent<br />

0.03 0.03<br />

= 0.05<br />

0.59<br />

0.8<br />

0.56 0.56<br />

= 0.95<br />

p(test indicates virus present) = 0.59 0.59<br />

(b)<br />

Figure 8.10 – (a) Revising the prior probabilities when the test indicates that the virus is<br />

present; (b) revising the prior probabilities when the test indicates that the virus is absent

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!