Continuous Distributions
Hey students! š Welcome to our exploration of continuous distributions - one of the most fascinating and practical topics in actuarial science! In this lesson, you'll discover how continuous probability distributions help us model real-world phenomena like insurance claims, stock prices, and even the time between customer arrivals at a bank. By the end of this lesson, you'll understand the four major continuous distributions (Uniform, Exponential, Gamma, and Normal), master their density and cumulative distribution functions, and learn powerful transformation techniques that actuaries use daily. Get ready to unlock the mathematical tools that help insurance companies calculate premiums and assess risks! šÆ
Understanding Continuous Distributions
Unlike discrete distributions that deal with countable outcomes (like rolling dice), continuous distributions handle variables that can take any value within a range. Think of measuring someone's height - they could be 5.7 feet, 5.73 feet, or 5.732 feet tall. There's no "next" possible height because we can always measure more precisely! š
The foundation of any continuous distribution lies in its probability density function (PDF), denoted as $f(x)$. Here's the key insight students: unlike discrete probabilities, the PDF doesn't give us the probability at a specific point (which is actually zero for continuous variables!). Instead, it tells us the relative likelihood of values occurring in that region.
The cumulative distribution function (CDF), written as $F(x)$, gives us the probability that our random variable $X$ is less than or equal to some value $x$. Mathematically, we express this as:
$$F(x) = P(X \leq x) = \int_{-\infty}^{x} f(t) dt$$
This relationship between PDF and CDF is crucial - the CDF is simply the area under the PDF curve from negative infinity up to point $x$! š
The Uniform Distribution
Let's start with the simplest continuous distribution - the Uniform distribution! Imagine you're waiting for a bus that arrives randomly between 10 AM and 11 AM. Every minute within that hour is equally likely for the bus to arrive. That's uniform distribution in action! š
The Uniform distribution over interval $[a, b]$ has a remarkably simple PDF:
$$f(x) = \begin{cases}
$\frac{1}{b-a}$ & \text{if } a $\leq$ x $\leq$ b \\
$0 & \text{otherwise}$
$\end{cases}$$$
The CDF is equally straightforward:
$$F(x) = \begin{cases}
0 & \text{if } x < a \\
$\frac{x-a}{b-a}$ & \text{if } a $\leq$ x $\leq$ b \\
1 & \text{if } x > b
$\end{cases}$$$
In actuarial science, uniform distributions often model scenarios where we have complete uncertainty within known bounds. For example, if an insurance company knows a claim will be settled sometime between $1,000 and $5,000, but has no reason to believe any amount is more likely than another, they might use a uniform distribution for initial modeling.
The mean of a uniform distribution is simply $\mu = \frac{a+b}{2}$ (the midpoint), and the variance is $\sigma^2 = \frac{(b-a)^2}{12}$.
The Exponential Distribution
Now let's explore the exponential distribution - a superstar in actuarial modeling! š This distribution is perfect for modeling "waiting times" or "time until failure." Think about how long your smartphone battery lasts, the time between insurance claims, or how long customers spend on hold with customer service.
The exponential distribution has a single parameter $\lambda > 0$ (called the rate parameter), and its PDF is:
$$f(x) = \lambda e^{-\lambda x} \text{ for } x \geq 0$$
The CDF has a beautifully simple form:
$$F(x) = 1 - e^{-\lambda x} \text{ for } x \geq 0$$
Here's what makes the exponential distribution special - it has the memoryless property! This means that if you've already waited 10 minutes for the bus, the probability of waiting another 5 minutes is the same as if you had just arrived at the bus stop. Mathematically: $P(X > s + t | X > s) = P(X > t)$.
Real-world example: Insurance companies use exponential distributions to model the time between claims for certain types of policies. If claims follow an exponential distribution with $\lambda = 0.1$ per month, then the average time between claims is $\frac{1}{\lambda} = 10$ months.
The mean is $\mu = \frac{1}{\lambda}$ and the variance is $\sigma^2 = \frac{1}{\lambda^2}$. Notice how the standard deviation equals the mean - that's a unique characteristic of exponential distributions!
The Gamma Distribution
The gamma distribution is like the exponential distribution's more flexible cousin! š While exponential distributions model the time until the first event, gamma distributions model the time until the $k$-th event occurs. It's defined by two parameters: shape parameter $\alpha > 0$ and rate parameter $\beta > 0$.
The PDF involves the gamma function $\Gamma(\alpha)$:
$$f(x) = \frac{\beta^\alpha}{\Gamma(\alpha)} x^{\alpha-1} e^{-\beta x} \text{ for } x > 0$$
Where $\Gamma(\alpha) = \int_0^\infty t^{\alpha-1} e^{-t} dt$. For positive integers, $\Gamma(n) = (n-1)!$.
The gamma distribution is incredibly versatile in actuarial applications. Insurance companies use it to model claim sizes, especially for property damage where smaller claims are more common than larger ones. When $\alpha = 1$, the gamma distribution becomes the exponential distribution - showing how these distributions are related!
Here's a fascinating fact: if you have $n$ independent exponential random variables with the same rate $\beta$, their sum follows a gamma distribution with parameters $\alpha = n$ and rate $\beta$. This makes gamma distributions perfect for modeling aggregate claims over time periods.
The mean is $\mu = \frac{\alpha}{\beta}$ and the variance is $\sigma^2 = \frac{\alpha}{\beta^2}$.
The Normal Distribution
Finally, we reach the crown jewel of continuous distributions - the Normal (or Gaussian) distribution! š This bell-shaped curve appears everywhere in nature and finance, from people's heights to stock market returns to measurement errors.
The normal distribution is characterized by two parameters: mean $\mu$ and standard deviation $\sigma > 0$. Its PDF is:
$$f(x) = \frac{1}{\sigma\sqrt{2\pi}} e^{-\frac{(x-\mu)^2}{2\sigma^2}}$$
Unfortunately, the CDF doesn't have a simple closed form, so we rely on numerical integration or standard normal tables.
The standard normal distribution (with $\mu = 0$ and $\sigma = 1$) is denoted as $Z \sim N(0,1)$, and any normal random variable can be standardized using: $Z = \frac{X - \mu}{\sigma}$.
In actuarial science, normal distributions model many phenomena due to the Central Limit Theorem. Even if individual claims aren't normally distributed, the average of many claims tends toward normality. This is why insurance companies can predict their total payouts with remarkable accuracy despite uncertainty in individual claims!
The famous 68-95-99.7 rule states that approximately 68% of values fall within one standard deviation of the mean, 95% within two standard deviations, and 99.7% within three standard deviations.
Transformation Techniques
Understanding how to transform distributions is crucial for actuarial work! š When you have a random variable $X$ with known distribution and create a new variable $Y = g(X)$ using some function $g$, you need to find the distribution of $Y$.
For a strictly increasing function $g$ with inverse $g^{-1}$, if $X$ has PDF $f_X(x)$, then $Y$ has PDF:
$$f_Y(y) = f_X(g^{-1}(y)) \cdot \left|\frac{d}{dy}g^{-1}(y)\right|$$
This technique is essential when modeling insurance scenarios where you need to transform claim amounts (perhaps adjusting for inflation) or when converting between different measurement scales.
Conclusion
We've journeyed through the four fundamental continuous distributions that form the backbone of actuarial science! The uniform distribution provides our simplest model for complete uncertainty within bounds, while the exponential distribution captures the essence of waiting times with its unique memoryless property. The gamma distribution extends this concept to model the time until multiple events, and the normal distribution gives us the versatile bell curve that appears throughout nature and finance. Together with transformation techniques, these tools enable actuaries to model complex real-world phenomena and make informed decisions about risk and uncertainty.
Study Notes
ā¢ Probability Density Function (PDF): $f(x)$ represents relative likelihood, not actual probability
ā¢ Cumulative Distribution Function (CDF): $F(x) = P(X \leq x) = \int_{-\infty}^{x} f(t) dt$
ā¢ Uniform Distribution: $f(x) = \frac{1}{b-a}$ for $x \in [a,b]$, $\mu = \frac{a+b}{2}$, $\sigma^2 = \frac{(b-a)^2}{12}$
ā¢ Exponential Distribution: $f(x) = \lambda e^{-\lambda x}$, $F(x) = 1 - e^{-\lambda x}$, $\mu = \frac{1}{\lambda}$, $\sigma^2 = \frac{1}{\lambda^2}$
ā¢ Memoryless Property: $P(X > s + t | X > s) = P(X > t)$ (exponential only)
ā¢ Gamma Distribution: $f(x) = \frac{\beta^\alpha}{\Gamma(\alpha)} x^{\alpha-1} e^{-\beta x}$, $\mu = \frac{\alpha}{\beta}$, $\sigma^2 = \frac{\alpha}{\beta^2}$
ā¢ Normal Distribution: $f(x) = \frac{1}{\sigma\sqrt{2\pi}} e^{-\frac{(x-\mu)^2}{2\sigma^2}}$
ā¢ Standardization: $Z = \frac{X - \mu}{\sigma}$ converts any normal to standard normal
ā¢ 68-95-99.7 Rule: 68% within 1Ļ, 95% within 2Ļ, 99.7% within 3Ļ of mean
ā¢ Transformation Formula: $f_Y(y) = f_X(g^{-1}(y)) \cdot \left|\frac{d}{dy}g^{-1}(y)\right|$