91Ó°ÊÓ

When dealing with binomial distributions, sometimes it can become cumbersome to compute probabilities directly, especially if the sample size (\( n \)) is large and the probability of success (\( p \)) is small. This is where the Poisson approximation comes in handy.
The Poisson distribution offers a simplified way to approximate the binomial distribution when two conditions are met:

• The number of trials (\( n \)) is large.
• The success probability (\( p \)) is small.

Under these conditions, we set our Poisson parameter, \( \lambda \), to be equal to \( np \). This parameter represents the expected number of successes in the number of trials.
In the airport metal detector example, calculating using the binomial distribution is impractical due to \( n = 500 \) and \( p = 0.005 \). Therefore, we approximate it with a Poisson distribution where \( \lambda = 500 \times 0.005 = 2.5 \). This approximation simplifies finding the probability of events significantly.

The Probability Mass Function (PMF) is a fundamental concept in probability theory used to describe the probabilities of different outcomes for discrete random variables.
For a Poisson distribution, the PMF is denoted as \( P(X=k) \) and given by the formula:\[ P(X=k) = \frac{e^{-\lambda} \lambda^k}{k!} \]where \( e \) is the base of the natural logarithm, \( \lambda \) is the average rate (\( np \) in Poisson approximation), and \( k \) represents the number of events (or successes) we are interested in. This formula gives the probability that a Poisson-distributed random variable \( X \) equals \( k \).
In the context of the metal detector, this translates to probabilities like \( P(X=5) \), which can be calculated using this formula, substituting \( k = 5 \) and \( \lambda = 2.5 \). The result reflects the probability of exactly 5 people activating the metal detector.

A Bernoulli trial is a basic experiment that results in a binary outcome: success or failure. Each trial is independent, adhering to its simplicity and ensuring that prior outcomes don't influence future ones. The probability of success is denoted by \( p \) while the failure is \( 1-p \).
In scenarios involving numerous Bernoulli trials, like evaluating if passengers activate an airport metal detector, the trials are discretely repeated:

• Each passenger passing through is a separate trial.
• "Activating the detector" is defined as the success, happening with a probability of \( 0.005 \).

Large numbers of trials with small probabilities collectively form a binomial distribution. However, when it's computationally feasible, especially with a large \( n \) as seen here, approximating the event counts with a Poisson distribution offers efficiency.
Thus, Bernoulli trials serve as the foundational building blocks, leading us towards complex distributions like binomial or their pragmatic cousin, the Poisson distribution.