Probability Summary Sheet Basic Probability Addition Rule ο· The addition rule provides a connection between the probabilities of two events, the probability of both occurring, and the probability of either occurring. o ππ π΄ ∪ π΅ = ππ π΄ + ππ π΅ − ππ(π΄ ∩ π΅) Karnaugh Maps ο· Karnaugh Maps summarise all the combinations of two events and their complements. π© π·π(π¨ ∩ π©) π·π(π¨′ ∩ π©) π·π(π©) π¨ π¨′ π©′ π·π(π¨ ∩ π©′) π·π(π¨′ ∩ π©′) π·π(π©′) π·π(π¨) π·π(π¨′ ) π. π Conditional Probability ο· The probability of A occurring given that B has occurred is represented by; o ππ π΄ π΅ = ππ(π΄∩π΅) ππ(π΅) Independent Events ο· If the probability of event A does not influence event B and vice versa, then; o β΅ ππ(π΄βπ΅) = Pr(π΄) o ∴ ππ π΄ ∩ π΅ = ππ π΄ × ππ π΅ Mutually Exclusive Events ο· If two events cannot occur simultaneously (mutually exclusive), then; o β΅ ππ(π΄ ∩ π΅) = 0 o ∴ ππ π΄ ∪ π΅ = ππ π΄ + ππ π΅ Combinations ο· Combinations refers to the number of ways n different objects can be arranged when taking r at a time, represented by π πΆπ where; o n Cr = n! n−r !r! Random Variables Definition of a Random Variable ο· A random variable is a function that assigns a numerical value to each outcome of an experiment. ο· A probability distribution consists of all values that a random variable can take, and their respective probabilities. ο· A probability distribution is described by a function p(x) where; o p x = Pr(X = x) o 0≤π π₯ ≤1 o π π₯ =1 Discrete Random Variables Discrete Random Variables ο· A random variable is said to be discrete if it can only assume a countable number of values. ο· These are typically represented by probability tables such as; X a b c d Pr(X=x) Pr(X=a) Pr(X=b) Pr(X=c) Pr(X=d) Expected Value of a Discrete Random Distribution ο· The expected value (denoted by E(X)) which is also known as the mean (µ) is given by; o E X = µ = x β Pr(X = x) where; ο§ E X + Y = E X + E(Y), where X and Y are discrete random variables ο§ πΈ[π π₯ ] = π(π₯) β π(π₯) ο§ E aX + b = aE X + b, where a and b are constants Median of a Random Distribution ο· The median of a random distribution is the value such that 50% of the distribution is greater than it, and 50% of the distribution is less than it. o The median (denoted by m) is calculated by adding the probabilities until such that Pr X ≤ m ≥ 0.5 π+π o If values exist such that Pr X ≤ a = 0.5, Pr X ≥ b = 0.5, then π = 2 Mode of a Random Distribution ο· The mode (denoted by M) of a random distribution is the most probable value of the random variable. o The mode is such that Pr(π = π) ≥ Pr π = π₯ , for all other values of x Variance and Standard Deviation ο· The variance and standard deviation are measures of spread – that is; how close to the mean the values of a random probability distribution are. ο· The variance of a random variable is defined by; o σ2 = Var X = E[(X − µ)2 ] = (π₯ − µ)2 β Pr(π = π₯) where; ο§ Var aX + b = a2 β Var(X), where a and b are constants ο· The standard deviation of a random variable is defined by; o π = ππ· π = πππ(π) Applications of Standard Deviation ο· A property of the standard deviation is such that the following ‘confidence intervals’ will apply o Pr µ − σ ≤ X ≤ µ + σ ≈ 0.68 o Pr µ − 2σ ≤ X ≤ µ + 2σ ≈ 0.95 o Pr µ − 3σ ≤ X ≤ µ + 3σ ≈ 0.997 Binomial Distributions Binomial Distributions ο· A binomial distribution (also referred to as a Bernoulli sequence) is a particular type of discrete probability distribution which possesses the following properties: o Each trial results in only one of two outcomes, one which is denoted as a success, or one which is denoted as a failure o The probability of success is constant for all trials o The trials are independent Binomial Random Variables ο· Binomial random variables are defined by; o ππ π = π₯ = n! π π₯ (1 − π)π −π₯ n−r !r! where; ο§ n: Number of trials ο§ p: Probability of success ο§ q: Probability of failure ο§ x: Number of successes ο· This can alternatively be expressed as X~Bi (n, p) Graphs of Binomial Distributions ο· The shape of a binomial distribution graph changes according to the values of n and p. ο· If p is changed; o When p<0.5, the graph is positively skewed. i.e. Lower values of x are favoured o When p=0.5, the graph is symmetrical o When p>0.5, the graph is negatively skewed. i.e. Higher values of x are favoured ο· If n is changed; o As n increases, the graph becomes increasingly ‘smooth’. o As n decreases, the graph becomes increasingly ‘rigid’. Mean for a Binomial Distribution ο· In binomial distributions, the mean is calculated by; ο· µ=E X =nβp ο· Note: The mode and median are not defined in binomial distributions. Spread for a Binomial Distribution ο· In binomial distributions, the variance is calculated by; o σ2 = Var X = n β p β 1 − p Markov Chains Markov Chains ο· Markov Chains possess the following properties; o The probability of a particular outcome is conditional only on the outcome before it o The conditional probabilities of each outcome are the same every time ο· Recurrence relationships can be set up to solve Markov chain questions. Examples of recurrence relationships are; o ai+1 = Pr ππ+1 ππ β ππ + Pr ππ+1 ππ β ππ o bi+1 = Pr ππ+1 ππ β ππ + Pr ππ+1 π β ππ ο· The long-term behaviour of Markov chains (also known as the steady state) can be modelled by assuming ai+1 = ai Continuous Random Variables Definition of a Continuous Random Variable ο· A continuous random variable is one that can take on any value in an interval of the real number system. ο· Due to the infinite number of decimal places to which a continuous random variable can be measured, a continuous random variable cannot take an exact value as it is rounded to the limits imposed by the method of measurement used. o Pr X = x = 0 for all values of x Continuous Distributions ο· Continuous distributions can be represented by a probability density functions (pdf). Probability density functions are functions that represent continuous distributions, and as such, possess the following properties; o f(x) ≥ 0 for all values of x π o ∫π f x dx = 1, where the domain of f(x) is (m,n) ο· Both of these properties must be proved to prove a function is probability density function ο· The probability of a continuous random variable falling between x=a and x=b is given by; o b Pr a ≤ X ≤ b = ∫a f x dx Mean for a Continuous Distribution ο· In a continuous distribution described by the probability density function with domain [a,b], the mean is calculated by; o b µ = E X = ∫a x β f x dx b E[g X ] = ∫a g(x) β f x dx ο§ Percentiles for a Continuous Distribution ο· In continuous distributions, the median is calculated by; o p n = ∫a x β f x dx , where p is the nth percentile To calculate the median, let p=0.5 To calculate the interquartile range (IQR), subtract the 75th percentile from the 25th percentile Mode for a Continuous Distribution ο· To calculate the mode; ο§ ο§ Differentiate the probability density function and find the probability values at the turning points o Find the probability values of the end-points of the probability density function. o Whichever of these has the maximum value of f(x) is the mode Spread for a Continuous Distribution ο· To calculate the variance, the following formula is used in continuous distributions; o o b b b σ2 = Var X = ∫a (x − µ)2 β f x dx = ∫a x 2 β f x dx − ∫a x β f x dx 2 Normal Distributions The Standard Normal Distribution ο· The standard normal distribution is a special type of continuous distribution with a probability density function which possesses the following properties; o It is defined by f x = 1 2 1 e−2 x 2π o It has a mean, median and mode of 0 o It has a standard deviation of 1 o It is an symmetrical around zero ο· The standard normal distribution can be represented as Z~N(0,1) Normal Distributions ο· A normal distribution is similar to the standard normal distribution, except it does not necessarily have a mean of 0 or standard deviation of 1. Its properties instead are; o Its median, mode and mean are the same value o Hence the value of the mean translates the graph horizontally o It is symmetrical about the mean o As the value of the standard deviation increases, the graph widens and vice versa ο· Normal distributions are represented by X~N(µ, σ2) ο· The graph of a normal distribution can be obtained by the following transformation upon the graph of the standard normal distribution; y o x, y → σx + μ, ο· σ This leads to the more general rule; o Pr X ≤ x = Pr Z ≤ x−μ σ