1. ELEC 303, Koushanfar, Fall’09 ELEC 303 – Random Signals Lecture 3 – Conditional probability Farinaz Koushanfar ECE Dept., Rice University Sept 1, 2009

2. ELEC 303, Koushanfar, Fall’09 Lecture outline Reading: Sections 1.4, 1.5 Review Independence – Event couple – Multiple events – Conditional – Independent Trials and Binomial

3. ELEC 303, Koushanfar, Fall’09 Bayes rule The total probability theorem is used in conjunction with the Bayes rule: Let A 1,A 2,…,An be disjoint events that form a partition of the sample space, and assume that P(A i ) > 0, for all i. Then, for any event B such that P(B)>0, we have

4. ELEC 303, Koushanfar, Fall’09 Bayes rule inference Figure courtesy of Bertsekas&Tsitsiklis, Introduction to Probability, 2008

5. ELEC 303, Koushanfar, Fall’09 The false positive puzzle Test for a certain rare disease is correct 95% of times, and if the person does not have the disease, the results are negative with prob.95 A random person drawn from the population has probability of.001 of having the disease Given that a person has just tested positive, what is the probability of having the disease?

6. ELEC 303, Koushanfar, Fall’09 Independence The occurrence of B does not change the probability of A, i.e., P(A|B)=P(A)  P(A  B)=P(A).P(B) Symmetric property Can we visualize this in terms of the sample space?

7. ELEC 303, Koushanfar, Fall’09 Independence? A and B disjoint  A and B c are disjoint P(A|B) = P(A|B c )

8. ELEC 303, Koushanfar, Fall’09 Effect of conditioning on independence Assume that A and B are independent If we are told that C occurred, are A and B independent? The events A and B are called conditionally independent if P(A  B|C)=P(A|C)P(B|C)

9. ELEC 303, Koushanfar, Fall’09 Conditional independence Independence is defined based on a probability law Conditional probability defines a probability law Example courtesy of Prof. Dahleh, MIT

10. ELEC 303, Koushanfar, Fall’09 Example Example: Rolling 2 dies, Let A and B are A={first roll is a 1}, B={second roll is a 4} C={The sum of the first and second rolls is 7} Is (A  B|C) independent?

11. ELEC 303, Koushanfar, Fall’09 Independence vs. pairwise independence Another example: 2 independent fair coin tosses A={first toss is H}, B={second toss is H} C={The two outcomes are different} P(C)=P(A)=P(B) P(C  A)=? P(C|A  B)=? Example courtesy of Prof. Dahleh, MIT

12. ELEC 303, Koushanfar, Fall’09 Independence of a collection of events Events A 1,A 2,…,A n are independent if For every subset S of {1,2,…,n} For 3 events, this amounts to 4 conditions: – P(A 1  A 2 ) – P(A 1  A 3 ) – P(A 2  A 3 ) – P(A 1  A 2  A 3 )

13. ELEC 303, Koushanfar, Fall’09 Reliability example: network connectivity P(series subsystem succeeds) P(parallel subsystem succeeds) Figure courtesy of Bertsekas&Tsitsiklis, Introduction to Probability, 2008

14. ELEC 303, Koushanfar, Fall’09 Independent trials and the Binomial probabilities Independent trials: sequence of independent but identical stages Bernoulli: there are 2 possible results at each stage Figure courtesy of Bertsekas&Tsitsiklis, Introduction to Probability, 2008

15. ELEC 303, Koushanfar, Fall’09 Bernoulli trial P(k)=P(k heads in an n-toss sequence) From the previous page, the probability of any given sequence having k heads: p k (1-p) n-k The total number of such sequences is Where we define the notation

16. ELEC 303, Koushanfar, Fall’09 Binomial formula The probabilities have to add up to 1!