Summary
Highlights
The video introduces Bayes' Theorem, emphasizing its derivation from the definition of conditional probability. It presents the formula for conditional probability as P(A|B) = P(A ∩ B) / P(B) and P(B|A) = P(B ∩ A) / P(A). By manipulating these equations, the video derives Bayes' formula: P(A|B) = P(B|A) * P(A) / P(B). The core idea is to find a conditional probability P(A|B) using available information where the events are flipped, P(B|A).
An example is presented: Joe, a randomly chosen individual from a large population where 3% are heroin users. A drug test correctly identifies users 95% of the time and non-users 90% of the time. The question is to find the probability that Joe is a heroin user given his test is positive.
The speaker writes down the given probabilities: P(Heroin User) = 0.03. Then, 'test correctly identifies users 95% of the time' is interpreted as P(Positive Test | Heroin User) = 0.95. Consequently, P(Negative Test | Heroin User) = 1 - 0.95 = 0.05. Similarly, 'correctly identifies non-users 90% of the time' is interpreted as P(Negative Test | Non-Heroin User) = 0.90, and P(Positive Test | Non-Heroin User) = 1 - 0.90 = 0.1.
To solve the problem, a tree diagram is used to visualize the probabilities. The first branches represent being a heroin user (0.03) or not a heroin user (0.97). From each of these, subsequent branches represent positive or negative test results based on the conditional probabilities derived earlier.
The problem asks for P(Heroin User | Positive Test). Using Bayes' formula: P(H|P) = P(P|H) * P(H) / P(P). The numerator, P(P|H) * P(H), is calculated as 0.95 * 0.03. For the denominator, P(P) (probability of a positive result), the two paths leading to a positive result in the tree diagram are summed: (0.03 * 0.95) + (0.97 * 0.1).
The calculation combines these values: P(H|P) = (0.03 * 0.95) / [(0.03 * 0.95) + (0.97 * 0.1)]. The final answer is 0.2271 (rounded to four decimal places).