Some more examples
+ Binomial: Given a population oF a very large lot and the probability of
defective items is p, what is the probability of getting x defective items in a
‘random sample of toms?
+ Hypergeometric: Given a finite lot of W items containing D defective items,
what is the probability of getting x defective items in a random sample of
items is drawn from the lot without replacement?
+ Poisson: Given the average number of defects observed on an item is A,
what is the probability of obtaining x defects on a randomly inspected
item?
Practice examples
1. production procsssoperstes witsaserancentoming (a0, boa
cap, Evey tour acarploaksoummc! prcuctetcken, Lexmnl (5°, 0-02)
Si the umber of nanconsrming te coumatase
2. 0.0 pulys ae proaucad using Lato #1, 40 are uk. fi
cel teapot Arann carpe stion —— Hyin. (700, 60 %)
prodicion parts cotanng x pais coming tom Labe #1
3. iret boards are produced ints of siz 20, The sample .
Si size Sls crawn fom the ot at one time and lasted. the jy/ur(fW'=20, O° 3, "=2)
iotcontains 3 ronconforming bosres and xa the numberof
Tenconocming boats ne sample
4. Let beth number of misprints none page ofa day, (j=)
rawapaper ihe average misprints pr page & 2 Suen
5, 1000 fi in pond, 100 of thm ae tagged. xis # of
tagged teh among S randomly caught sh
Some examples
8. Acids in abutting ate assumed to occu randomly (
‘san average rte a 9 per jen. Tho ul be x
ents te coring AD |
1 Reece 0 pam 2 argent wx Hy
trrorpages nero selection of pages d
8. The proebiy theta eaesman wl make a ale on one oe
Talis Each coy. thesaosran mate Oeale Lax Setown'al (19, 0-7)
ena the rue of ates ade one dy.
8. The average rumbor a fowe sera YSO1®
np eter fee OS eaten meow tow (100 )
Fire 1dr ol ota abi {
10, The probably tat a baseball slayer wil make aoe 4)
Inew 07, Lat xdorote the numer of ree vests he Bnomal ( 7,07)
‘wilmake ina game of seven free throw attempts,Normal distribution
+ Measure the length of the bolt and draw a histogram
fle)
Normal distribution
+ Probability calculations:
o( xa) [tio = vormef (04.9, poe)
+ What is the probability that x = a?
flies)>a
Standard normal distribution
+ Consider the transformation (standardization):
xoN ly *) ze Xk
E (é) 20 Vou(2’)
(‘ )
P(X #28) - *(2>-8)
2 i Plesas) = $25)
= PGs) = 09937
Example
+ The time to resolve customer complaints is a critical quality
characteristic for many organizations. Suppose that this time in
a financial organization, say, x—is normally distributed with
mean y: = 40 hours and standard deviation o = 2 hours. What is
the probability that a customer complaint will be resolved in
greater than 35 hours?
| 340
-»(:< 85")
=1-0(-2 (1-425)
ple > 35
ve
= (2.5)Example
Example
Alinkage is made of three shafts with length normally distributed as’
1N(75,0.09) —-N(25,0.25) (60,016)
1) Whats the distribution of the linkage? 0 s)
2) Suppose your customer won't accepta linkage if itis longer
than 160.5 cm. How much percentage of your linkages will be
rejected by the customer? What if your customer won't accept
a linkage if it is longer than 159.5 cm?
ea aExample
3) Suppose you are looking for a customer to sell your linkages
to, Each potential customer has an upper specification limit
(USL) for the linkages he can accept, meaning that he will
reject a linkage if itis longer than USL.. You want to work with
a customer who will accept at least 95% of your linkages.
What USL shall the customer have in order for you to be
willing to work with him?
PC acceptn'y e dsbese | y 0-45
s Pl L< vs) xo
, Example
P(g ¢ bach) 70-7
P(z < Usicp) 9 °F
Wt-B y (re)
=
USL & $ (ot) x0 + fe
zp psy fo5+4 [ho
z
160 18
¢(- gu lon (2 bee
\
)
fCentral limit theorem
* Cental Xw
ut a7 a "
Central limit theorem
+ Central limit theorem: if x,,.x2,...,x, are independent random
variables with means j. and variance a?, then as n tends to 0
the sample mean
emean
“¥ N(Pe)
et
pury ech yibadions a
Note: regardless ofthe underyng probably density function ofthe _
Population data, the distbuton of te means of samples rom the =x,
Population will be normally distributedCentral limit theorem
+ Most physical, chemical, ... phenomena that we observe, such
as temperature, hardness often fit a normal distribution as they
are hypothesized to be a sum of large number of independent
random effects.
+ These random effects may not be normally distributed, but
when the sum (or mean) is observed, it behaves like a normal
distribution, a direct consequence of central limit theorem
Example: Binomial to Normal
5 of theme
fe prob of socen
eB of saeco
no dow
Bocemiol —» Novmal(nf ; bl -p)
Lognormal distribution
. } 52)
Za N(8,% )
exp{z) ~ dagntoeal of faibotoars
2
and
L 15) exh (40-4002 )x loxp(B\-1)
2(2)- txf(o-+.u?) bin (2) = exp C0) (f(D)Chi-square distribution rs
4%), ) kn n(o) L Aid |
2
: 2
Ye MEQ tt ky 9 Ye X@
J
lew oegreg Joue)em
™
Ym. (y)= 20
Student t-distribution
xv w(o,t) yo Ky
Xn, £-dertri betoin ”
Wh
F-distribution .
Xv Xn You Km