1. Home
  2. >Statistics
  3. >Discrete Probabilities
Help

Negative binomial distribution

Get the most by viewing this topic in your current grade. Pick your course now.

?
Intros
Lessons

  1. • Deriving negative binomial distribution
    • Formula for negative binomial distribution
    • Relation of geometric distribution to the negative binomial distribution
?
Examples
Lessons
  1. Identifying Negative Binomial Distributions
    Identify which of the following experiments below are negative binomial distributions?
    i.
    A fair coin is flipped until head comes up 4 times. What is the probability that the coin will be flipped exactly 6 times?
    ii.
    Cards are drawn out of a deck until 2 exactly aces are drawn. What is the probability that a total of 10 cards will be drawn?
    iii.
    An urn contains 3 red balls and 2 black balls. If 2 balls are drawn with replacement what is the probability that 1 of them will be black?
    iv.
    Roll a die until the first six comes up. What is the probability that this will take 3 rolls?
    1. Determining the Negative Binomial Distribution
      A fair coin is flipped until head comes up 4 times. What is the probability that the coin will be flipped exactly 6 times?
      1. Determining the Cumulative Negative Binomial Distribution
        A sculptor is making 3 exhibits for an art gallery. There is a probability of 0.75 that every piece of wood she carves into will be good enough to be part of the exhibit. What is the probability that she uses 4 pieces of wood or less?
        Topic Notes
        ?

        Introduction to Negative Binomial Distribution

        The negative binomial distribution is a crucial concept in probability theory and statistics. Our introduction video provides a comprehensive overview, making it an essential starting point for understanding this distribution. The negative binomial distribution is closely related to both the geometric distribution and the binomial distribution, sharing similarities and key differences with each. It models the number of successes in a sequence of independent Bernoulli trials before a specified number of failures occur. This distribution is particularly useful in various fields, including biology, finance, and quality control. Unlike the binomial distribution, which focuses on a fixed number of trials, the negative binomial distribution allows for a variable number of trials until a certain number of failures is reached. This flexibility makes it ideal for modeling scenarios where events continue until a specific condition is met, such as the number of sales calls needed to achieve a target number of successful sales.

        Understanding the Negative Binomial Distribution

        The negative binomial distribution is a powerful statistical concept that extends the principles of the geometric distribution to account for multiple successes. This distribution is particularly useful in modeling scenarios where we're interested in the number of trials needed to achieve a specific number of successes, rather than just one success as in the geometric distribution.

        To understand the negative binomial distribution, let's first revisit its simpler counterpart, the geometric distribution. The geometric distribution models the number of trials needed to achieve the first success in a series of independent Bernoulli trials. For example, it could represent the number of coin flips needed to get the first heads.

        Now, imagine we want to extend this concept to multiple successes. This is where the negative binomial distribution comes into play. Instead of stopping at the first success, we continue until we reach a predetermined number of successes. This makes the negative binomial distribution more versatile and applicable to a wider range of real-world scenarios.

        Let's illustrate this with a coin flipping example. Suppose we're flipping a fair coin and we want to know the probability of getting 3 heads (our successes) in 10 flips. This scenario perfectly fits the negative binomial distribution model. We're not just interested in the first heads, but in achieving a specific number of heads (3) within a certain number of trials (10).

        The formula for the negative binomial distribution probability mass function is:

        P(X = k) = C(k-1, r-1) * p^r * (1-p)^(k-r)

        Where:

        • k is the number of trials
        • r is the number of successes
        • p is the probability of success on each trial
        • C(k-1, r-1) is the binomial coefficient, also known as "k-1 choose r-1"

        Let's break down each component of this formula:

        1. C(k-1, r-1): This represents the number of ways to choose r-1 successes from k-1 trials. It accounts for the different ways the successes can be arranged within the trials.

        2. p^r: This term represents the probability of achieving r successes.

        3. (1-p)^(k-r): This represents the probability of failing in the remaining k-r trials.

        The negative binomial distribution formula essentially combines these probabilities to give us the likelihood of achieving our desired number of successes in a specific number of trials.

        Comparing this to the geometric distribution, we can see that the negative binomial distribution is indeed an extension. The geometric distribution is actually a special case of the negative binomial distribution where r = 1, meaning we're only interested in the first success.

        The flexibility of the negative binomial distribution makes it invaluable in various fields. In biology, it can model the distribution of parasites among hosts. In business, it can predict the number of sales calls needed to close a certain number of deals. In quality control, it can estimate the number of items that need to be inspected to find a specific number of defects.

        Understanding the negative binomial distribution and its relationship to the geometric distribution provides a powerful tool for analyzing and predicting outcomes in scenarios involving multiple successes. By grasping this concept, you'll be better equipped to model and interpret a wide range of real-world phenomena that involve repeated trials and multiple desired outcomes.

        Negative Binomial Distribution Formula and Its Components

        The negative binomial distribution is a powerful statistical tool used to model the number of successes in a sequence of independent Bernoulli trials before a specified number of failures occur. At the heart of this distribution lies its formula, which we'll break down and explain in detail.

        The negative binomial distribution formula is:

        P(X = x) = C(x + r - 1, x) * p^x * (1-p)^r

        Let's examine each component of this formula:

        1. P(X = x): This represents the probability of achieving x successes before the r-th failure occurs.

        2. C(x + r - 1, x): This is the binomial coefficient, also known as "x + r - 1 choose x". It calculates the number of ways to arrange x successes and r - 1 failures in any order.

        3. p^x: This term represents the probability of x successes occurring, where p is the probability of success on each trial.

        4. (1-p)^r: This term represents the probability of r failures occurring, where 1-p is the probability of failure on each trial.

        Now, let's discuss the significance of the key parameters:

        n (number of trials): In the negative binomial distribution, n is not fixed. The number of trials continues until a specified number of failures (r) is reached. This makes it different from the binomial distribution, where n is predetermined.

        x (number of successes): This represents the number of successful outcomes before reaching the specified number of failures. It's the variable we're typically interested in predicting or analyzing.

        p (probability of success): This is the probability of success on each individual trial. It remains constant throughout the sequence of trials.

        r (number of failures): This is the predetermined number of failures that will end the sequence of trials. It's a fixed parameter in the negative binomial distribution.

        To use the negative binomial distribution formula, follow these steps:

        1. Determine the values of x, r, and p for your specific scenario.

        2. Calculate the binomial coefficient C(x + r - 1, x).

        3. Compute p^x and (1-p)^r.

        4. Multiply all these terms together to get the final probability.

        Let's walk through an example:

        Suppose we're flipping a coin (p = 0.5) and want to know the probability of getting 3 heads (x = 3) before getting 2 tails (r = 2).

        Step 1: We have x = 3, r = 2, and p = 0.5

        Step 2: C(3 + 2 - 1, 3) = C(4, 3) = 4

        Step 3: 0.5^3 = 0.125 and (1-0.5)^2 = 0.25

        Step 4: 4 * 0.125 * 0.25 = 0.125

        Therefore, the probability of getting 3 heads before 2 tails when flipping a fair coin is 0.125 or 12.5%.

        The negative binomial distribution has numerous real-world applications. In quality control, it can model the number of items inspected before finding a certain number of defects. In epidemiology, it can represent the number of people who must be vaccinated to prevent a specific number of disease cases. In finance, it can model the number of trades before a certain number of losses occur.

        Understanding the components of the negative binomial distribution formula allows for more accurate modeling of scenarios where we're interested in the number of successes before a certain number of failures. By manipulating the parameters r and p, we

        Comparing Binomial and Negative Binomial Distributions

        Understanding the differences and similarities between binomial distribution models and negative binomial distributions is crucial for statisticians, data scientists, and researchers across various fields. Both distributions are discrete probability distributions that deal with the number of successes in a series of independent trials, but they have distinct characteristics and applications.

        Binomial Distribution

        The binomial distribution models the number of successes in a fixed number of independent Bernoulli trials. Each trial has only two possible outcomes: success or failure. The probability of success remains constant for each trial. Key characteristics of the binomial distribution include:

        • Fixed number of trials (n)
        • Constant probability of success (p) for each trial
        • Independent trials
        • Interest in the number of successes (X)

        Negative Binomial Distribution

        The negative binomial distribution, on the other hand, models the number of failures before a specified number of successes occurs. It does not have a fixed number of trials but continues until a predetermined number of successes is achieved. Key characteristics include:

        • Variable number of trials
        • Constant probability of success (p) for each trial
        • Independent trials
        • Interest in the number of failures before reaching a specified number of successes (r)

        Similarities

        Despite their differences, binomial and negative binomial distributions share some similarities:

        • Both are discrete probability distributions
        • Both involve a series of independent Bernoulli trials
        • The probability of success (p) remains constant for each trial in both distributions
        • Both can be used to model events with two possible outcomes

        When to Use Each Distribution

        The choice between binomial and negative binomial distributions depends on the nature of the problem and the information available:

        • Use the binomial distribution when:
          • The number of trials is fixed
          • You're interested in the number of successes
          • Each trial has a constant probability of success
        • Use the negative binomial distribution when:
          • The number of trials is not fixed
          • You're interested in the number of failures before a specific number of successes
          • Each trial has a constant probability of success

        Real-World Applications

        Binomial distribution applications:

        • Quality control: Determining the number of defective items in a batch
        • Medical trials: Assessing the number of patients responding to a treatment
        • Marketing: Analyzing the number of successful sales calls out of a fixed number of attempts
        • Elections: Predicting the number of votes a candidate might receive

        Negative binomial distribution applications:

        • Customer acquisition: Modeling the number of sales calls needed to acquire a certain number of customers
        • Epidemiology: Studying the number of disease-free days before a specified number of infections occur
        • Manufacturing: Analyzing the number of units produced before achieving a target number of high-quality items
        • Sports analytics: Modeling the number of at-bats before a baseball player hits a certain number of home runs

        Comparison Table

        Examples of Negative Binomial Distribution

        The negative binomial distribution is a versatile probability distribution that models the number of successes in a sequence of independent Bernoulli trials before a specified number of failures occur. Let's explore several examples with step-by-step solutions to demonstrate its application in various contexts.

        Example 1: Quality Control in Manufacturing

        A factory produces electronic components with a 90% success rate. The quality control team wants to know the probability of finding the 5th defective component after inspecting 50 components.

        Step 1: Identify the parameters
        p (probability of success) = 0.90
        q (probability of failure) = 1 - p = 0.10
        r (number of failures) = 5
        x (total number of trials) = 50

        Step 2: Apply the negative binomial formula
        P(X = 50) = C(49, 4) * (0.90^45) * (0.10^5)

        Step 3: Calculate the result
        P(X = 50) 0.1328 or 13.28%

        Interpretation: There is a 13.28% chance of finding the 5th defective component on the 50th inspection.

        Example 2: Marketing Campaign Success

        A marketing team launches a new campaign and considers it successful after acquiring 10 new customers. If the probability of converting a lead is 20%, what is the probability of achieving success after contacting 40 potential customers?

        Step 1: Identify the parameters
        p (probability of success) = 0.20
        q (probability of failure) = 1 - p = 0.80
        r (number of successes) = 10
        x (total number of trials) = 40

        Step 2: Apply the negative binomial formula
        P(X = 40) = C(39, 9) * (0.20^10) * (0.80^30)

        Step 3: Calculate the result
        P(X = 40) 0.0915 or 9.15%

        Interpretation: There is a 9.15% chance of acquiring the 10th new customer on the 40th contact.

        Example 3: Epidemiology Study

        In a study of a rare disease, researchers want to find 5 cases before concluding their investigation. If the prevalence of the disease is 1%, what is the probability of finding the 5th case after screening 400 individuals?

        Step 1: Identify the parameters
        p (probability of success) = 0.01
        q (probability of failure) = 1 - p = 0.99
        r (number of successes) = 5
        x (total number of trials) = 400

        Step 2: Apply the negative binomial formula
        P(X = 400) = C(399, 4) * (0.01^5) * (0.99^395)

        Step 3: Calculate the result
        P(X = 400) 0.0366 or 3.66%

        Interpretation: There is a 3.66% chance of finding the 5th case of the rare disease on the 400th individual screened.

        Example 4: Sports Analytics

        A basketball player has a 70% free throw success rate. What is the probability that he will make his 5th successful free throw on his 8th attempt?

        Step 1: Identify the parameters

        Relationship Between Geometric and Negative Binomial Distributions

        The geometric distribution and the negative binomial distribution are closely related probability distributions in statistics. In fact, the geometric distribution is a special case of the negative binomial distribution. This relationship is fundamental to understanding both distributions and their applications in various fields, including quality control, reliability analysis, and risk assessment.

        To understand this relationship, let's first review the definitions of both distributions. The negative binomial distribution models the number of failures before a specified number of successes occur in a sequence of independent Bernoulli trials. On the other hand, the geometric distribution specifically models the number of failures before the first success occurs.

        The key to understanding how the geometric distribution is a special case of the negative binomial distribution lies in the parameter that represents the number of successes. In the negative binomial distribution, this parameter (often denoted as r) can be any positive integer. However, when we set r = 1, we are essentially looking at the number of failures before the first success, which is precisely what the geometric distribution describes.

        Let's derive the geometric distribution formula from the negative binomial formula to illustrate this relationship mathematically. The probability mass function (PMF) of the negative binomial distribution is given by:

        P(X = k) = C(k + r - 1, r - 1) * p^r * (1-p)^k

        Where:

        • k is the number of failures
        • r is the number of successes
        • p is the probability of success on each trial
        • C(n,k) represents the binomial coefficient

        Now, let's set r = 1 to derive the geometric distribution:

        P(X = k) = C(k + 1 - 1, 1 - 1) * p^1 * (1-p)^k

        Simplifying:

        P(X = k) = C(k, 0) * p * (1-p)^k

        The binomial coefficient C(k, 0) is always equal to 1, so we can further simplify:

        P(X = k) = p * (1-p)^k

        This final form is the well-known probability mass function of the geometric distribution. This derivation clearly demonstrates how the geometric distribution emerges as a special case of the negative binomial distribution when we consider only the first success (r = 1).

        To illustrate this relationship with an example, let's consider a quality control scenario in a manufacturing process. Suppose we're interested in the number of defective items produced before a non-defective item is observed.

        If we use the negative binomial distribution with r = 1 and p = 0.8 (assuming an 80% chance of producing a non-defective item), we get:

        P(X = k) = C(k + 1 - 1, 1 - 1) * 0.8^1 * (1-0.8)^k = 0.8 * 0.2^k

        This is identical to the geometric distribution formula for the same scenario:

        P(X = k) = 0.8 * (1-0.8)^k = 0.8 * 0.2^k

        Both distributions would give the same probabilities for various numbers of defective items before the first non-defective item is produced.

        Understanding this relationship has practical implications. It allows statisticians and data scientists to use the more general negative binomial distribution in situations where they might have previously used only the geometric distribution. This flexibility can be particularly useful when analyzing processes where multiple successes are of interest, not just the first one.

        Moreover, software packages and statistical tools that implement the negative binomial distribution can be used to calculate geometric distribution probabilities by simply setting the number of successes to 1. This relationship also

        Conclusion and Further Applications

        The negative binomial distribution is a powerful statistical tool for modeling count data with overdispersion. As demonstrated in the introduction video, it extends the Poisson distribution by allowing for greater variance. Key points include its use in modeling the number of failures before a specified number of successes, its application in various fields such as epidemiology and ecology, and its flexibility in handling clustered data. The video provides a crucial foundation for understanding this distribution's mechanics and practical applications. To deepen your knowledge, explore advanced topics like zero-inflated negative binomial models, mixture models, and Bayesian approaches to negative binomial regression. Further applications can be found in areas like risk analysis, quality control, and marketing research. By mastering the Poisson distribution, you'll enhance your ability to analyze and interpret complex count data across diverse disciplines.

        FAQs

        Here are some frequently asked questions about the negative binomial distribution:

        1. What is a negative binomial distribution?

        A negative binomial distribution is a discrete probability distribution that models the number of successes in a sequence of independent Bernoulli trials before a specified number of failures occur. It is an extension of the geometric distribution and is useful for modeling count data with overdispersion.

        2. How do you write a negative binomial distribution?

        The probability mass function of a negative binomial distribution is written as:
        P(X = k) = C(k + r - 1, r - 1) * p^r * (1-p)^k
        Where k is the number of failures, r is the number of successes, p is the probability of success on each trial, and C(n,k) is the binomial coefficient.

        3. What is the rule of negative binomial distribution?

        The negative binomial distribution follows these rules: 1. There is a fixed number of desired successes (r). 2. Trials continue until the r-th success is observed. 3. Each trial is independent. 4. The probability of success (p) remains constant for all trials. 5. The random variable X represents the number of failures before the r-th success.

        4. When would you use a negative binomial distribution?

        You would use a negative binomial distribution when: 1. You're modeling the number of failures before a specified number of successes. 2. The number of trials is not fixed. 3. You're dealing with count data that shows overdispersion (variance greater than the mean). 4. In scenarios like modeling the number of sales calls before achieving a target number of sales, or the number of disease-free days before a certain number of infections occur.

        5. What is the difference between binomial and negative binomial distributions?

        The main differences are: 1. Binomial distribution has a fixed number of trials, while negative binomial has a variable number of trials. 2. Binomial distribution counts the number of successes in a fixed number of trials, while negative binomial counts the number of failures before a fixed number of successes. 3. Binomial distribution stops after a predetermined number of trials, while negative binomial stops after reaching a predetermined number of successes.

        Prerequisite Topics for Understanding Negative Binomial Distribution

        When delving into the world of probability and statistics, it's crucial to build a strong foundation before tackling more complex concepts. The negative binomial distribution is an advanced topic that requires a solid understanding of several prerequisite subjects. By mastering these fundamental concepts, students can better grasp the intricacies of the negative binomial distribution and its applications in various fields.

        One of the most important prerequisites is the geometric distribution. This probability distribution models the number of trials needed to achieve the first success in a series of independent Bernoulli trials. Understanding the geometric distribution formula is essential because the negative binomial distribution can be viewed as an extension of this concept. It builds upon the idea of waiting for a specific number of successes rather than just the first one.

        Another crucial prerequisite is the binomial distribution. This distribution describes the number of successes in a fixed number of independent Bernoulli trials. Familiarity with binomial distribution models is vital because the negative binomial distribution shares similarities in its structure and application. Both distributions deal with discrete events and involve a series of trials, but they differ in their stopping conditions.

        The Poisson distribution is also an important prerequisite topic. While it may not seem directly related at first glance, understanding the Poisson distribution helps in grasping the concept of rare events and their occurrence over time or space. This knowledge is valuable when working with the negative binomial distribution, especially in scenarios where it's used to model overdispersed count data.

        By thoroughly studying these prerequisite topics, students can develop a comprehensive understanding of probability distributions and their interrelationships. This knowledge serves as a strong foundation for exploring the negative binomial distribution. For instance, recognizing the similarities and differences between the geometric and negative binomial distributions allows for a deeper appreciation of how the latter extends the concept to multiple successes.

        Moreover, the ability to compare and contrast binomial and negative binomial distributions enhances one's analytical skills in choosing the appropriate model for different real-world scenarios. Understanding the Poisson distribution's properties also aids in recognizing situations where the negative binomial distribution might be a more suitable alternative for modeling count data with greater variability.

        In conclusion, mastering these prerequisite topics is not just about memorizing formulas or concepts. It's about building a interconnected web of knowledge that allows for a more intuitive and comprehensive understanding of the negative binomial distribution. This solid foundation will not only make learning the negative binomial distribution easier but will also enhance overall statistical reasoning and problem-solving skills in various academic and professional contexts.

        • Negative Binomial Distribution: P(n)=(n−1)C(x−1)px(1−p)n−xP(n)=_{(n-1)}C_{(x-1)}p^x(1-p)^{n-x}
        nn: number of trials
        xx: number of success in n trials
        pp: probability of success in each trial
        P(n)P(n): probability of getting the xx success on the nthn^{th} trial
        Characteristic