Probability Distributions for Continuous Random Variables: The Uniform Distribution (Sabit Olas¬l¬kl¬Da¼ g¬l¬m) The uniform distribution is a probability distribution that has equal probabilities for all possible outcomes of the random variable f(x) Total area under the uniform probability density function is 1.0 xmax x xmin 1 Ozan Eksi, TOBB-ETU It has propability density function of 9 8 1 < for xmin < x < xmax = f (x) = xmax xmin ; : 0 elsewhere –Its mean is = E(X) = Z xmax xmin 2 xmax + xmin xf (x)dx = 2 Ozan Eksi, TOBB-ETU It has propability density function of 9 8 1 < for xmin < x < xmax = f (x) = xmax xmin ; : 0 elsewhere –Its variance is 2 = E[(X Z xmax = xmin )2 ] (x 1 2 ) f (x)dx = (xmax 12 3 xmin)2 Ozan Eksi, TOBB-ETU Ex: 2 x 6 aras¬nda tan¬mlanm¬ş sabit olas¬l¬kl¬ da¼ g¬l¬m düşünelim. Bunun olas¬l¬k fonksiyonu 1 2 x 6 için f (x) = = 0:25 6 2 ki şu şekilde gösterilebilir f(x) .25 2 6 4 x Ozan Eksi, TOBB-ETU Ortalamas¬: xmax + xmin 2 + 6 = = =4 2 2 Varyasyonu: 2 = 1 (x max 12 1 2 xmin) = (6 12 5 2)2 = 1:333 Ozan Eksi, TOBB-ETU The Normal Distribution (Normal Da¼ g¬l¬m) The normal distribution is the most important distribution in the statistical theory –It is bell-shaped –It is symmetrical around the mean –Its mean, median and mode are equal –Location is determined by the mean, –Spread is determined by the standard deviation, 6 Ozan Eksi, TOBB-ETU –The random variable has an in…nite theoretical range: 1 to +1 f(x) s µ x Mean = Median = Mode By varying the parameters µ and s , we obtain different normal distributions 7 Ozan Eksi, TOBB-ETU 8 Ozan Eksi, TOBB-ETU A (normal) random variable is the one having normal distribution where the probability density function is 1x ( )2 1 f (x) = p e 2 for 1 < x < 1; > 0 2 The normal distribution closely approximates the probability distributions of a wide range of random variables 9 Ozan Eksi, TOBB-ETU Computations of probabilities are direct and elegant Distributions of sample means approach a normal distribution given a “large”sample size* If random variable X has a normal distribution with and variance 2 , then it is shown as X N ( ; 2) 10 Ozan Eksi, TOBB-ETU Cumulative Normal Distribution: When X N ( ; 2), cumulative distribution function is Z x0 F (x0) = P (X x0) = f (x)dx 1 f(x) 0 x x0 11 Ozan Eksi, TOBB-ETU The probability for a range of values is measured by the area under the curve P (a < X < b) = F (b) a µ 12 b F (a) x Ozan Eksi, TOBB-ETU The total area under the curve is 1.0, and the curve is symmetric, so half is above the mean, half is below f(X) P( −∞ < X < μ) = 0.5 0.5 P(μ < X < ∞ ) = 0.5 0.5 µ X P(−∞ < X < ∞) = 1.0 13 Ozan Eksi, TOBB-ETU The Standardized Normal (Standart Normal Da¼ g¬l¬m) Any normal distribution (with any mean and variance combination) can be transformed into the standardized normal distribution (Z), with mean 0 and variance 1 14 Ozan Eksi, TOBB-ETU This need to transform X units into Z units by subtracting the mean of X and dividing by its standard deviation X Z= It obtains the following f(Z) Z ~ N(0,1) 1 0 15 Z Ozan Eksi, TOBB-ETU Ex: E¼ ger X ortalamas¬100, standart sapmas¬50 olan rassal bir de¼ gişken ise, X = 200 de¼ gerinin Z karş¬l¬g¼¬şudur X 200 100 Z= = =2 50 Buna göre X = 200 de¼ geri X de¼ gişkeninin ortalamas¬olan 100’den 2 standart sapma yüksektedir. Böylece X=200 de¼ gerinin, X’in alabilece¼ gi tüm de¼ gerlere için göreli yerini bulmuş oluruz 16 Ozan Eksi, TOBB-ETU Note that the distribution is the same, only the scale is standardized b −μ a −μ P(a < X < b) = P <Z< σ σ b −μ a −μ = F − F σ σ f(x) a µ b x a −μ σ 0 b −μ σ Z 17 Ozan Eksi, TOBB-ETU The Standardized Normal Table gives probability for any value of z –Normal da¼ g¬l¬ma sahip bir X rassal de¼ gişkeni için P(a < X < b) de¼ gerini bulal¬m Önce X’in a ve b’ye eşit oldu¼ gu de¼ gerleri Z’ye çevirebilir, sonra da kümülatif normal tablosunu kullanabiliriz –Ex: X ortalamas¬8.0, standart sapmas¬5 olan normal da¼ g¬l¬ma sahip bir rassal de¼ gişken olsun (yani X N (8; 25)). P(X < 8.6) de¼ gerini 18 Ozan Eksi, TOBB-ETU bulal¬m X Z= = 8:6 8 5 = 0:12, P (Z < 0:12) = 0:547 µ=8 s = 10 8 8.6 µ=0 s =1 X P(X < 8.6) 0 0.12 Z P(Z < 0.12) –Yani X rassal de¼ gişkeninin alabilece¼ gi de¼ gerlerin %54.78’i 8.6’n¬n alt¬ndad¬r 19 Ozan Eksi, TOBB-ETU 20 Ozan Eksi, TOBB-ETU For upper tail (üst kuyruk) properties –Ex: P (Z > 2:00) =? P (Z < 2:00) = 0:9772 ) P (Z > 2:00) = 1 21 0:9772 = 0:0228 Ozan Eksi, TOBB-ETU For negative Z-values, use the fact that it is symmetric distribution Ex: P (Z < ) 2:00) =? P (Z < 2:00) = 0:9772 P (Z < 2:00) = 1 0:9772 = 0:0228 .9772 .9772 .0228 .0228 Z 22 Z Ozan Eksi, TOBB-ETU Ex: Finding the X value for a Known Probability –X N (8; 25) ise X’in hangi de¼ geri X’in alabilece¼ gi tüm de¼ gerlerin %20’sinin üstündedir? .80 .20 ? 8.0 -0.84 0 23 Ozan Eksi, TOBB-ETU Z de¼ geri için bahsi geçen de¼ gerin 0.84 oldu¼ gunu standart normal tablosundan biliyoruz. O halde X Z= ) X= + Z = 8 + ( 0:84)5 = 3:8 24 Ozan Eksi, TOBB-ETU Ex: Araba yedek parças¬üreten bir şirketin üretti¼ gi bir ürünün dayan¬m süresi normal da¼ g¬l¬ma sahiptir ve ortalamas¬1,250 hafta, standart sapmas¬ da 250 haftad¬r. Bu ürünlerden rastgele seçilen bir tanesinin 900 ila 1,300 hafta aras¬nda dayanma olas¬l¬g¼¬nedir? 25 Ozan Eksi, TOBB-ETU P (900 < X < 1300) = P ( 900 <Z< 1300 900 1250 1300 1250 = P( <Z< ) 250 250 = P ( 1:2 < Z < 0:2) = F (0:2) F ( 1:2) = 0:5793 (1 0:8643) = 0:44 26 Ozan Eksi, TOBB-ETU ) Assessing Normality: Not all continuous random variables are normally distributed. It is important to evaluate how well the data is approximated by a normal distribution. However, there are tests that can be applied, for instance, by the use of statistical programs 27 Ozan Eksi, TOBB-ETU