28
Statistical Statistical Distributions Distributions

Statistical Distribution

Embed Size (px)

DESCRIPTION

Statics

Citation preview

Statistical Statistical DistributionsDistributions

Uniform DistributionA R.V. is uniformly distributed on the

interval (a,b) if it probability functionFully defined by (a,b)

P(x) = 1/(b-a) for a <= x <= b = 0 otherwise

Uniform Distribution Probability Function

1 10

1

1/9

Probability that x is between 2 and 7.5? Probability that x = 8?

1 10

1

1/9

Uniform DistributionThe cumulative distribution of a

uniform RV is

F(x) = 0 for x < a = (x-a)/(b-a) for a <= x <= b

= 1 otherwise

Uniform Distribution Cumulative Function

1 10

1

Uniform DistributionDiscrete vs. Continuous• Discrete RV

– Number showing on a die• Continuous RV

– Time of arrival – When programming, make it discrete

to some number of decimal places

Uniform Distribution• Mean = (a+b)/2• Variance = (b-a)2 /12

• P (x < X < y) = F (y) – F (x)= (y-a) - (x-a) = y – x – a + a = y - x b-a b-a b – a b-a

Uniform - ExampleA bus arrives at a bus stop every 20

minutes starting at 6:40 until 8:40. A passenger does not know the schedule but randomly arrives between 7:00 and 7:30 every morning. What is the probability the passenger waits more than 5 minutes.

Uniform Solution

5 10 15 20 25 30 40

1

1/30

X = RV, Uniform (0,30) -- i.e. 7:00 – 7:30Bus: 7:00, 7:20, 7:40Yellow Box <= 5 minute wait

A B C

P (x > 5) = A + C = 1 – B = 5/6

Arithmetic MeanGiven a set of measurements y1,

y2, y3,… yn

Mean = (y1+y2+…yn) / n

VarianceVariance of a set of measurements

y1, y2, y3,… yn is the average of the deviations of the measurements about their mean (m).

V = σ2 = (1/n) Σ (yi – m)2i=1..n

Variance ExampleYi= 12, 10, 9, 8, 14, 7, 15, 6, 14, 10m = 10.5V= σ2 = (1/10) ((12-10.5)2 + (10-10.5)2 +

…. = (1/10) (1.52 + .52 + 1.52….) = (1/10) (88.5)

= 8.85Standard Deviation = σ = 2.975

Normal Distribution• Has 2 parameters

–Mean - μ–Variance – σ2

–Also, Standard deviation - σ

Normal Dist.

0-3 -2 -1 1 2 3

Mean +- n σ

.3413 .1359

.0215 .0013

Normal Distribution• Standard Normal Distribution has

– Mean = 0 StdDev = 1• Convert non-standard to standard

to use the tablesZ value = # of StdDev from the meanZ is value used for reading table Z = (x – m)

σ

Normal - ExampleThe scores on a college entrance exam

are normally distributed with a mean of 75 and a standard deviation of 10. What % of scores fall between 70 & 90?

Z(70) = (70 – 75)/10 = - 0.5Z(90) = (90 – 75)/10 = 1.5.6915 - .5 = .1915 + .9332 - .5 = .4332

= .6247 or 62.47%

Exponential Distribution

A RV X is exponentially distributed with parameter > 0 if probability function

Mean = 1/Variance = 1 / 2 e = 2.71828182

e xP(x) =

For x >= 0

= 0 Otherwise

Exponential Distribution

• Often used to model interarrival times when arrivals are random and those which are highly variable.

• In these instances lambda is a rate– e.g. Arrivals or services per hour

• Also models catastrophic component failure, e.g. light bulbs burning out

Exponential Rates• Engine fails every 3000 hours

– Mean: Average lifetime is 3000 hours– = 1/3000 = 0.00033333

• Arrivals are 5 every hour– Mean: Interarrival time is 12 minutes– = 1 / 5 = 0.2

• Mean = 1 /

Exponential Distribution

Probability Function

x

f(x)

See handout for various graphs.

Exponential Distribution

Cumulative FunctionGiven Mean = 1/ Variance = 1/ 2

F(x) = P (X <=x) = 1 – e - x

Exponential Distribution

Cumulative Function (<=)

x

1

F(x)

Forgetfulness PropertyGiven: the occurrence of events conforms

to an exponential distribution:The probability of an event in the next x-

unit time frame is independent on the time since the last event.

That is, the behavior during the next x-units of time is independent upon the behavior during the past y-units of time.

Forgetfulness Example

• The lifetime of an electrical component is exponentially distributed with a mean of .

• What does this mean??

Forgetfulness Examples

The following all have the same probability• Probability that a new component lasts

the first 1000 hours.• Probability that a component lasts the

next 1000 hours given that it has been working for 2500 hours.

• Probability that a component lasts the next 1000 hours given that I have no idea how long it has been working.

Solution to Example• Suppose the mean lifetime of

the component is 3000 hours.• = 1/3000• P(X >= 1000) = 1 – P(X <=

1000) 1 – (1-e -1/3* 1) = e -1/3 = .717

How do we apply these?

1. We may be given the information that events occur according to a known distribution.

2. We may collect data and must determine if it conforms to a known distribution.