STATISTICAL METHODS FOR BUSINESS UNIT 2: RANDOM VARIABLES 2.1.- Random variable. Discrete and continuous variables 2.2.- Probability distribution of a random variable 2.3.- Characteristics of a random variable. Expected value and variance 2.4.- Tchebyshev’s inequality Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business UNIT 2. GOALS • To intuitively understand the concept of random variable and its relevance in the field of economics. • To analyse discrete and continuous random variables. • To compute cumulative probabilities and probabilities of intervals. • To be able to compute and interpret the expectation and the variance of a random variable. • To apply Chebyshev’s Inequality, and to understand its relevance in applications. Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business STATISTICAL METHODS FOR BUSINESS UNIT 2: RANDOM VARIABLES 2.1.- Random variables. Discrete and continuous variables Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business EMPIRICAL STUDIES Under certainty • No. of branches of a bank. • Last year’s profits. Under uncertainty • Employment level for next year. • Inflation expected for next month. Future events Statistical variables Non-exhaustive analyses Random variables Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business RANDOM VARIABLES Description of a random variable and its probability distribution X “Daily demand of oil stations in a town (thousand litres)” Probability calculus: P(X12), P(6<X15) Summary of X (statistical measures) Description of the probability distribution of X: Usual probability models Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Random variables 1 1 Working (W) 1 P(W) P(X = 1) RV (X) P(NW) P(X = 0) Not working (NW) 0 0 0 Probability Random experiment Outcome Induced probability Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Classes of random variables No. of employees in a shop. No. of customers in a bank office. Population employed in a sector. DISCRETE [Finite or countably infinite range] CONTINUOUS National income of a country. Inflation level. Consumption of oil. [Non-countable range] Waiting time in a semaphore. Earnings in a lottery. Tariff of a service HYBRID ( ) ( Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business STATISTICAL VARIABLES •Discrete •Continuous •Hybrid Values + Frequencies Mean Variance RANDOM VARIABLES ALEATORIAS VARIABLES CLASS DESCRIPTION CHARACTERISTICS •Discrete •Continuous •Hybrid Values + Probabilities Expectation Variance Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business STATISTICAL METHODS FOR BUSINESS UNIT 2: RANDOM VARIABLES 2.2.- Probability distribution of a random variable Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Discrete random variables ➨ Labor status X discrete RV Not working X=0 P(X=0) = 0.25 Working X=1 P(X=1) = 0.75 Probability function P(x) P : x ∈ℜ → [ 0, 1 ] P( x)≥0 ∑ P ( x i )=1 0.75 0.25 0 i 1 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Variable aleatoria continua Continuous random variables X “Profits of a bank’s branch” p1 p2 L0 L1 L2 L3 L4 L5 ... L k X continuous Probability density Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Probability density p1 p2 L0 L1 L2 L3 L4 L5 ... Lk Density = Intervals of smaller length L0 L1 Probability on the interval Length of the interval ... L k Density function f(x) x Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business •FUNCIÓN DE DENSIDAD Density function + f : x ∈ℜ → ℜ f ( x)≥0 +∞ ∫−∞ f ( x) dx=1 Total area = 1 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business DISTRIBUTION FUNCTION Definition Meaning F : x∈ℜ → F ( x) F ( x)= P( X ≤x)∈ [ 0,1 ] Probability accumulated up to value x. Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Cumulative probability V.A. DISCRETA Not working ➨ X=0 P(X=0) = 0.25 Working ➨ X=1 P(X=1) = 0.75 0 1 X { 0 if x<0, F ( x )= 0. 25 if 0≤x<1, 1 if x≥1 . i Distribution function F ( xi )= ∑ P( x j ) j=1 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business •Nº MEDALLAS OBTENIDAS POR UN ATLETA X “No. of medals won by an athlete” P(x) Probability function 0,5 Point probability 0,2 0,1 0 F(x) 1 2 3 Distribution function 1 0,9 Cumulative probability 0,7 0,5 0 1 2 3 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business DISTRIBUTION FUNCTION OF A CONTINUOUS VARIABLE x F ( x)=∫−∞ f (t ) dt Area = Probability accumulated up to x x Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Graphical representations 1 1 F(x) 0 X discrete F(x) X continuous Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Distribution function Properties 1. F non-decreasing monotonic: 2. x 1 <x 2 ⇒ F ( x 1 )≤ F( x 2 ) Lim x→−∞ F( x)= 0 3. Lim x→+∞ F( x)= 1 4. F is right continuous: Lim + h→0 F( x+h )= F( x ) Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Distribution function Probability function DISCRETE Density function CONTINUOUS Random variable Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Relationship between probability function and distribution function (discrete RVs) ✒ ✒ i F ( xi )= ∑ P( x j ) j=1 P ( X = x i )=F ( x i )−F ( xi−1 ) Relationship between density function and distribution function (continuous RVs) ✒ ✒ x F ( x)=∫−∞ f (t ) dt f (x)=F '(x) Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business •Ilustración function Probability •P(x) •P(x) •0.5 P ( X = x i )=F ( x i )−F ( xi−1 ) •0.5 •0.2 •0.1 •0 •1 •2 i F ( xi )= ∑ P( x j ) j=1 •3 •0.2 •0.1 •0 •1 •2 •3 •1 •2 •3 Distribution function •F(x) •F(x) •1 •0. 9 •0. •1 •0. 9 •0. 7 •0. 5 7 •0. 5 •0 •1 •2 •3 •0 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Probability of an interval PROBABILIDADES P ( a< X ≤b )=P ( X ≤b )− P ( X ≤a )= F ( b )−F ( a ) Continuous case b P (a< X ≤b)=∫a f ( x) dx Area = Probability of the interval •a •b Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Change of variable X Y = g(X) X discrete Y discrete X continuous Y discrete X continuous Y continuous Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business STATISTICAL METHODS FOR BUSINESS UNIT 2: RANDOM VARIABLES 2.3.- Characteristics of a random variable. Expectation and variance Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Characteristics of a random variable VARIABLES RANDOM STATISTICAL Value Frequency x1 F(x) X f1 k x2 f2 ̄x =∑ x f i i Synthesis i=1 E ( X )= μ ..... xk fk 2 S =( X −̄x ) 2 Variance 2 2 σ =E [ X −E ( X )] Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Expectation Deviation Random variable X Random error X- Expected value E(X)= E[X- ]=0 The expected deviation around is null. Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Expected value X discrete E ( X )=∫ x dF( x) X continuous ℜ +∞ E ( X )=∑ x i pi E( X )= ∫ x f ( x) dx i −∞ Properties • E(c) = c a , c∈ℜ • E(aX) = a E(X) • E(X+c) = E(X) + c • E(X+Y) = E(X)+E(Y) Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Variance Var ( X )=E [ ( X − μ ) X DISCRETE 2 ] 2 Var ( X )=∑ ( x i − μ ) pi i +∞ X CONTINUOUS 2 Var ( X )= ∫ ( x− μ ) f ( x ) dx −∞ Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Properties of the variance Shortened formula Var( X )= E ( X ) − μ • 2 0 2 2 a,c∈ℜ • Var(X+c) = Var(X) • Var(aX) = a2 Var(X) Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business STATISTICAL METHODS FOR BUSINESS UNIT 2: RANDOM VARIABLES 2.4.- Tchebyshev’s inequality Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business TCHEBYSHEV’S INEQUALITY ∣X −μ∣≥ε μ−ε μ ∣X −μ∣<ε μ+ε Let X be a RV with finite expectation and variance. Then for any 2 positive constant : and σ P (∣ X − μ∣≥ε )≤ 2 ε 2 σ P (∣ X − μ∣<ε )≥1− 2 ε Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business Tchebyshev’s bounds Value k Lower bound for P(|X-E(X)|<k) Upper bound for P(|X-E(X)|k) 1 0 1 2 0.75 0.25 3 0.89 0.11 4 0.9375 0.0625 5 0.96 0.04 10 0.99 0.01 Universidad de Oviedo. Facultad de Economía y Empresa. Grado en ADE. Statistical Methods for Business