91Ó°ÊÓ

The binomial distribution is a fundamental concept in probability and statistics, frequently used to model situations with two possible outcomes. Think of it as flipping a coin a certain number of times and counting the number of heads.
In the context of our problem, we have a diagnostic test that can either result in a positive or negative outcome for people known not to have a disease. The binomial distribution becomes handy because each test is independent, with a fixed probability of yielding a positive result, denoted by the parameter \( p \).

• First, we have \( n \) trials (in our example, the number of tests conducted, such as 20).
• Each trial results in a success (positive test) with probability \( p \) and a failure (negative test) with probability \( 1-p \).
• The number of successes (false positives) is represented by \( X \).

The probability of observing \( x \) successes (false positives) in \( n \) trials is given by the probability mass function:\[ P(X=x) = \binom{n}{x} p^x (1-p)^{n-x} \]where \( \binom{n}{x} \) is the binomial coefficient, calculated as \( \frac{n!}{x!(n-x)!} \). This expression captures the essence of the binomial setting in which each test outcome is influenced by the probability \( p \).

An unbiased estimator is a statistic used to estimate a parameter of a population. It is termed 'unbiased' if, on average, it hits the true parameter it's estimating.
In simpler terms, if you repeatedly calculated this estimator using different samples from the population, the average of these calculations would converge to the actual parameter value.
For our binomial example:

• The maximum likelihood estimator (MLE) of \( p \) is \( \hat{p} = \frac{X}{n} \).
• This representation uses the ratio of observed successes (false positives) to total trials (tests).
• The expected value of \( \hat{p} \) is calculated as:\[ E(\hat{p}) = E\left(\frac{X}{n}\right) = \frac{1}{n}E(X) \]
• Given that \( E(X) = np \) for a binomial distribution, it follows:\[ E(\hat{p}) = \frac{np}{n} = p \]

This shows that \( \hat{p} = \frac{X}{n} \) is unbiased, because its expected value equals the true parameter \( p \). This ensures that we aren't consistently overestimating or underestimating \( p \).

The probability mass function (PMF) is a mathematical function that provides the probability of each possible value of a discrete random variable. It's a helpful tool for understanding and quantifying random processes involving distinct outcomes.
For the binomial distribution identified in this problem:

• The PMF helps us to pinpoint the likelihood of getting exactly \( x \) false positives in \( n \) tests.
• Expressed formulaically, the PMF of a binomial distribution is:\[ P(X=x) = \binom{n}{x} p^x (1-p)^{n-x} \]
• This function includes:
  • \( \binom{n}{x} \): the number of ways to choose \( x \) successes out of \( n \) trials.
  • \( p^x \): the probability of realizing exactly \( x \) successes.
  • \( (1-p)^{n-x} \): the probability of \( n-x \) failures.

By using the PMF, one can calculate and understand the distribution of outcomes across different scenarios, such as the number of false positives when a test is applied to healthy individuals. This foundational understanding aids in estimating parameters accurately, such as using the PMF to determine the likelihood for observed data, as seen in our exercise.