This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Section 1.5 Other Useful Continuous Distributions . Many data sets in practice have histograms which have shapes unimodal, skewed, rises to maximum and then decreases. Some distributions having these properties are: 1. Lognormal Distribution Definition : The variable x is said to said to follow LN ( μ, σ ) if its density is ( 29 , 2 1 2 ) ln( 2 2 1 x e x x μ σ π ) ,  ( σ μ x f = otherwise , . 1 Facts: (i) Putting y = ln (x), it is easy to check ∫ ∞ ∞ = 1, ) ,  ( dy y f σ μ for all μ ε R, σ > 0. (ii) If x ~ L N ( μ, σ ), then y = ln (x) ~ N ( μ, σ ). This fact is used to compute proportions. Example 1. The variable x = modulus of elasticity (MOE) of Wood Joist Floor Systems with Creep is observed, in a study, to follow LN (.375, .25). Find (i) Proportion of systems with x < 3. (ii) Proportion of systems with 1 < x < 2. (iii) For what value of c, only 1% of systems have x (MOE) > c. That is, c is nothing but the 99 th percentile) Solution . (i) Proportion of systems with x < 3. ⇒ y = ln (x) < ln (3) = 1.098 Now y ~ N (.375, .25) Therefore, y < 1.0908 ⇔ .25 .375 1.098 25 . 375 . < = y z i.e., z ~ N (0, 1) and Propo. of z < 2.88 From normal tables, proportion of (z < 2.88) = .998. 2 Hence, prop. of systems with x < 3 = .998 (or 99.8%). (ii) Similarly, Prop. of systems with 1 < x < 2 ) 2 ( ln ) ln( ) 1 ( ln < < ⇔ x ) 2 ( ln ) 1 ( ln < < ⇔ y 25 . 375 . ) 2 ln( 25 . 375 . 25 . 375 . ) 1 ln( < < ⇔ y z ~ n (0,1); proportion of z with 1.512 < z < 1.272 Prop. of z < 1.272 = .8984 (see tables) Prof. of z < 1.512 = .0653 Prop. of z : 1.512 < z < 1.272: .8331 (subtract) Prop. of x with 1 < x < 2 = 83.3% (iii) Need to find c 220d prof x > c = .01 ⇔ prop. of x with x ≤ c = .99 Therefore, find c 220d .99 25 . 375 . ) ln( 25 . 375 . ) ln( = ≤ c x i.e., c is 220d proportion of 99 . 25 . 375 . ) ln( z = = ≤ c From tables, 99 th percentile of z is 2.33. 3 Therefore, 2.33 25 . 375 . ) ln( = c Solving ln ( c ) = .9575 ⇒ c = 2.605. 2. Weibull Distribution: W (α , β) This is an extension of exponential distribution. Definition : A continuous variable x follows W (α , β) if its density is , 1  x x e x α α β α β α ) ,  ( β α x f = elsewhere. , (i) Different values of α and β give positively and negatively skewed distributions. 4 Some graphs of graphs of Weibull distributions are: 5 (ii) Prop. of x with x < t ∫ = t dx x f ) ,  ( β α ∫  = t x dx x e 1 α β α α β α ∫  = α β t y dy e ....
View
Full Document
 Summer '08
 Palaniappan
 Poisson Distribution, Probability theory, Binomial distribution, Discrete probability distribution

Click to edit the document details