Probability Ridho Rahmadi Machine Learning Informatika, FTI UII March 2019 Think about an experiment from which outcome is unpredictable; but suppose you know the set of possible outcomes (sample space S). (a) S = {girl, boy } (b) S = {2, 3, 5, 4, 6, 7, 1} (c) S = (0, ∞) (d) S = {head, tail} Sample space and events Any subset E of the sample space S is called an event I An event consists of possible outcomes. I If the outcome is contained in E , then the event E has occurred. Sample space and events For example, 1. Based on Figure (a), E = {girl} 2. Based on Figure (b), E = {all outcomes in S starting with a 7} 3. Based on Figure (c), E = (0, 2) 4. Based on Figure (d), E = {tail} From 1 to 4, E is the event that the child is a girl, E is the event that the horse number 7 wins, E is the event that a patient gets an injection of dosage less than 2, and E is the event of the coin falling with tail on top, respectively. For any two events E and F of a sample space S, I If E = {girl} and F = {boy }, then E ∪ F = {girl, boy }. The E ∪ F occurs if either E or F occur. I If E = (0, 5) and F = (2, 10), then EF (denoting the intersection between the two) occurs if only both E and F occur, e.g., EF = (2, 5). Note that () above indicates interval. 0 I For any event E , E is a complement of E , consisting all outcome in S that are not in E . For example, if E = {girl}, 0 E = {boy }. I If, based on Figure (b), E = {an outcome started with 5}, F = {an outcome ended with 5}, then EF = ∅ (null event or event with no outcome), which means E and F cannot occur and thus E and F are mutually exclusive. Axioms of probability I 0 ≤ P(E ) ≤ 1 I P(S) = 1 I For any sequence of mutually exclusive events E1 , E2 , . . . , En , i.e.,SEi Ej = ∅, i 6= j, P( ni=1 ) = Σni=1 , n = 1, 2, , . . . , ∞. I 1 = P(S) = P(E ∪ E c ) = P(E ) + P(E c ), equivalently P(E c ) = 1 − P(E ). I P(E ∪ F ) = P(E ) + P(E ) − P(EF ) Odds of an event The odds of an event is defined by P(A) P(A) = P(Ac ) 1 − P(A) For example, if P(A) = 3/4, then the odds is 3, which means that the event A likely occurs 3 times higher than it does not. Conditional probability Given, S = {(i, j)}, i = j = 1, 2, . . . , 6, where i = the first die lands on side i, and j = the second die lands on side j. What is the probability that the sum of rolling two dice = 8, given that the first die landed on side 3? Answer: (3, 1), (3, 2), (3, 3), (3, 4), (3, 5), (3, 6) From the six possibilities of two dice given the first die landed on side 3, only (3, 5) that gives 8 as a total sum (1 out of six possibilities). Thus, the conditional probability is 1/6. Conditional probability Probability of event E , given that event F occurred, P(E |F ) = , only when P(F ) > 0. P(EF ) P(F )