91Ó°ÊÓ

The Negative Binomial Distribution is a fascinating type of probability distribution that's perfect for certain experiments, especially those dealing with sequential trials. Imagine opening a sequence of oysters to find a specific number of pearls.
This scenario perfectly fits the negative binomial distribution.

• The key here is counting how many trials (or in our case, failures) you need until you hit a specific number of successes.
• In our oyster example, a success means finding a pearl.
• The probability of finding a pearl in any single oyster is given by \( p \).

The formula for the probability distribution is:\[ \mathbf{P}(X = r) = \binom{r+k}{r} (1-p)^r p^k \]This equation calculates the probability of finding exactly \( r \) oysters without pearls before successfully collecting \( k \) pearls.
This is essentially the definition of the negative binomial distribution, encapsulating both the successes needed and the failures encountered in achieving those successes.
Moreover, this distribution is known for its flexibility in summing up to one, validating its nature as a probability distribution. This means: \[ \sum_{r=0}^{\infty} \binom{r+k}{r} (1-p)^r p^k = 1 \]

Understanding the mean and variance of our distribution gives us insights into the average and variability of our oyster trials.
Let's break down these components:

• The **mean** \( E(X) \) of a negative binomial distribution measures the ballpark number of failures (oysters without pearls) expected before finding \( k \) pearls. The formula is:\[ E(X) = \frac{k(1-p)}{p} \]This is calculated by considering the number of failures and the probability associated with each failure.
• The **variance** \( \text{Var}(X) \) tells us about the spread of these outcomes, essentially showing how much variation exists:\[ \text{Var}(X) = \frac{k(1-p)}{p^2} \]Variance is larger when you expect more uncertainty about the number of failures.
  The expressions for mean and variance both rely on the probability \( p \) of an oyster containing a pearl and the requirement to find exactly \( k \) pearls, addressing both the expected value and the unpredictability inherent in the process.

When dealing with limits, we explore what happens as our scenario grows infinitely large. In this problem, we consider what happens as \( k \), the number of pearls needed, approaches infinity.
Here's where things get interesting:

• We modify our probability of success \( p \) by setting it as \( p = 1 - \lambda / k \).
• As \( k \) grows, \( \lambda / k \) diminishes, effectively altering the scenario's scope.
  This shift impacts the calculations, leading us towards a new distribution.

By substitution, as \( k \to \infty \), the negative binomial distribution converges to a **Poisson distribution**. The Poisson distribution is characterized by the parameter \( \lambda \), which emerges due to this limit process.
This transition from a negative binomial to Poisson distribution is a fascinating demonstration of limits in the probability distributions. It showcases how probability behavior can simplify under certain conditions, making it easier to handle large-scale problems.