0% found this document useful (0 votes)
13 views7 pagesLecture 01 (3.1.25) - Merge-Invert
Hello
Uploaded by
meena.iitd.684Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
13 views7 pagesLecture 01 (3.1.25) - Merge-Invert
Hello
Uploaded by
meena.iitd.684Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 7
Lecture 1 (3.1.25)
nhoduction +p Statistics
Randomness -
Statistical Randomness. A numewe sequence which caains 10 Vecagnrrable
pattern
Probal Theo: is a mothematicad fiamework thot allows Us to desonbe
ond analyse relidom phenomenon. 2+ 8 a Ingle oF uncertainty
— sample space: +" of an experiment is the set of ott possible ovtvanes
oF an experiment
Event: ‘E° i8 a subset of sample space -%, and we say thot ‘6' has ocured
iF the actual outcome fs in
eq “Bssing a cin ee LH, th
TBssing two coins: (HH, Ht, TH, TT
Selecting a number G0M(0,)): -%= (0,1)
No: oF“ covip-14 infected ppt: St = Lo,1,2,
=~ max. populotiony
=? Naive definition of _prebobility
PE) ~ 11 = No. of outcomes fayoumble +o &
ol “Toto no: of ovtcomes: in SL
Le
SH 15 a finite set
AU outcomes are assumed 4 be quatly Uloly
= Moltipticotion Rule.
Consider a compound experiment consisting two sub-experiments. A ond 8.
Sippose Exp-A has ‘a! possible outcomes and Exp. 6 has ‘6’ possible
ovtcomes fer each oF the ‘a’ outcomes, then the compound experiment has ‘ab’
possible out comes:
Gg- Tssing a coin and rolling a dice
Sa {ty e+ {4,2,3,4,5,6h
Laws THU, #2, H3,HU, HS, HO. TL, 12, 73,TU TS, 76H
= [nol = (nal= 6, |Rmel= 12 = [Rergl= |Aal-\fel
— Bxioms of Proto
o> For ony event E. P@)=O
Ba
b> ROD)
cr foy disjoint events &. and £2,
PEUE.) = PE)+ P(ED
(= Poof: 2 f.--3
D= AVAL ANA >
wed= IAL = 1AUAsI = Air ipy 1Al> a2) = P(AD+ PiA2)
1.21 La pe] Let
3S PlAuBY= PIA + PCB) - RCA NB)
G Prot: = f-.-3
A=f--y Bh--3
(ALB) = ACA) + n@)- (Ane) --— Principle oF inclusion anc Exclusion
m(pug) = n(ave) = nA)+ n(g)-nfans) = 2) + n(@) - naned
nD NCA) 1) ny ®
FA) © e@)- Plane) Lecture 2 (7.1.25)
> Sathdoy Problem
G] Let There be én’ people In a group. Whot © the probability that cHleart ro
oF them hove same. brvthoLays Cuor-Leap year)
Total - no ppl have same birthday
365" = 8G cn eo!
es 565"
— Lemma: Let E be an event
PE\+ pet) = 4
Proof ENE =
P(Eve) = PE). pe) = Porr= 2
@) Whot shold be the min: value of n fer whidh PCED = os
SCAM S 4 Nz 28
365" 2 ~
— Lemma: let €:,€2..--- En 2 vis igsn
Then (Oe)
6 Eres
PaooF Axiom 3 and "for
— Rrioms of proven bly ore cansistent with naive definition
Rmark: Aromake definition of prtabclity [ more generat t naive definition.
Eg» “Tasting. a coin ee {uty Not consistent
(POD = pu , p= 06 with naive deF®
—7 Lemma: For naive def , every singleton subset of 2 has prbebitity Vi
=? Collection of events: Cotfechion of al substts of - (Power set of 2).
Benoted by A
= {ity = A= (OLE ter
Probability is a fonction fom @ — [2.11
+ Fidd/rigerra © Feld, probablity meosure Lecture 3 (8.1.25)
— Geomehic _Probabi ity
Def” . Suppose for an event € , favourable area is ar(E) and the total oven of
sample space ax). then the geometric frobability is given os: | (PIE)= ax(E)
aren?
Remark: “This definihon can be. extended ty ony dimension
consider an cquiloterot +viange thot is inscribed in the civele - A chord of the cinta
is chosen ot mondom , hot is the probobility thot the chow’ '= Ranger ‘thon Ae
side of the ‘viongle. —_ ferrtraad's Pondolox
e Favourable, 3
a] oe le. g a 3 Fosgate
PE= = + (Pee) aa Pey= we- 4
wo 8 4 +, OF
— Measure Theoretic. Framework :
Ler 2 be a non-empty sat Ler A te a callection of sutsets of 2 satisfying
ey mee
by 3F AS A then ATS &
cy AREA then AUB SA
Eg O m= tard, Ge £6. ther ort3}
f= {e023
DeF® {4.2} S known as a triviol field
@ Me faved} , t= {% *, 3, {bet}. — Ferd
Lemena: for a finite %, iF 181>2.3 9 fed ubch is nota hivial Feld a power wt
Proof: WwlOG: Let = {O,..--.. Am}
Corsiden ‘the cotlection & = {6 2. for, A/fou)
< show thet Fis a fed
Lemma: [et fr ve a fieth on % op Ai.----, AN =. then AU ALUA3..UMe
Prof By definition.
emma: Let & bea fiell on anh A. P2= # then ANAS &
Proof: (AM Aas = ALU AL
fan Ay = ©
_ ues &
@nnays &
A. NAa S &
Remark: O Zp the definition of field, union con be replaced with intersection
® In athe definition oF field, in the third poh, we can place
9- a=
€= One
Def: c- Felok: Let & be a collection oF subsets of U satisfying
OQ ase
@® sr AS then A Sar
@ FF ALA He me ahen OA SO
Lemma: Every = Fell is a frets Lecture 4 (10.1.25)
— 2 iS a non-empty set @ is a collection of subsets of 2
Feld: Coe) - Pros said be o elo if it sobsfes
O ner
© af AE then ASS &
@ af ABE A shen AUB Em
Eg- 22 @0 = fx: o< net, x€R}
IR denotes the set of real nos.
E = QNO0= {xz ecx Probability meosure: let GX,A) be a measurable space. A set function P? A— R
15 said tp be a probability measure. 09 (S%, A) iF iF satshes
@ MM 20 yvAew
© P(r =14
© For, Ar, Pe, EA such that AINA)= 6 ¥ixy then
POM) = Ze)
=? Definition. (L.A. ®) is known os & probability space
= Lerma QNoive definition oF probability sotisfes the definition of probability measure
@ AU fields ave not o- Fields. © Chink oF an example)
Fay Consider +vssing of two coins together
soe fin, we bn oh
Bm = 2 (Power set of -2)
ler the naive probability ve the probability mesure
(im) 5 a prota tly space Mi
PQ)= 0, POV=) | KHy)= 1/4
= Discrete _Probobi uf Space
TF iS of countable set and (2, &, 1) bea probability space Then (2,6. F)
is called oO. discover probability space:
Q =
Eq: s
o Gl8
PUY) = 2/2° for iEN
Rn = 2 Ru) = 2
iN) =¢ viej
Lemma: For a disorele probability space defining measure for singteton subset of 2 is
enough to define the probability measure:
Proof Hw
Ceroma . Let C2, &,P) be % prob space SF ABE A and ASB the Pi = PH
Proof: B= Ay (B\A). and AN (B\A)= >. Gs ASB)
- P= P(A) + P@\ad
PB\AY = P18)- PA) ZO
- Rf) = e@)
Lemma: Let As, Pa. -- =, An ER
“then (UA) < = P(A)
Proofs Huo
Lemma : Let Ps, Aa,----- 6
then, PC Ui) = 2 P(A)
Henk: He.
—+ Bord «- fielh: Let 2 be a@ Non-emply sot The © Field genenatedt by open subsets of
“A is known as Boek e-Feld on
Remar: For reo 0-5, open intenvats ave “the open subsets. 5> the o oleira
erated open intervals on IR Cor any subset of R) is said we be He
oorspandl J Borel «- field 3
Remark: for probability measves, we witt te considering Smet «- frelds only Lecture 5 (15.1.25)
— Remask OFF is a countable set , we wil usually toke A = 2%, je: the
assouvted o-Fdd is the power set
© Gre o-fidd on Ris the e-Febll generated by open inienvals in
Def: Genoroted o- field: Let -% be a non-empty set anol © be a collection of
subsets_of 1%. Then the o-fieth generated by C is definect
os @§ (@ = Ne st Awa Feld ond #26
Sometimes, = () is refered as minimal o-fietch
contoining &
Remark» Any Borel set (any member of Borel s-fied) on FR can be represented os a
countable union and intersection Of open intewals in R
@ Zs {1} a Borel set on R? j Every singleton set 1s a
Hint) 1) (I Mn. it /n) Borel! ito R
=? Random voriable (RV): lef -& be a non-empty set and & be a o- alge on 2.
A rondom voriobte ts a fonchon x: A—R se,
2B) = fad: x@e BYE Mm, ¥ BERR), Where BCA) isa
Gorel «algebra on ®
Let C2, A) be a menswable space and AEM Then Ty (@) = {28 weA
is a random wriable ©) otherwise.
— Bearer % 18 a eV on CAA) HR x '(f-00,x]) € A for oll ER
=7 A (G0) = [o FF x <0
AS iF ocxct
if x3!
© algebra generated by A is {4-2 .A, Ary
& aCe.) = fw: 1) =23
= {2 1W) Properties of distribvhen fincton
— Theorem: Let C2%,6-,1P) be a prob: space and x be an RV on Gn, A,P)
farther let dist: find of x be Fx Then-
© FrE~) = © @® Fx i8 @ monotonically increpsing finct
® FeGor=2 © Peis a right -contrnvovs ‘fonet
— Theorem: 9 funchion F: &— Cod sotsfying the above four properPes is a
dist func” of an Rv.
Rint: (R, BOR), PL) XC) = 9, (R, B(@),R> , obseme PF Py
=> De Support of an RY:
“he support set of on Rv x is clenoed by the set Sx and defined as
the minima set st Pees. =2
Remore: Fox any A2Sx , PxEM=2
Eg. Consider the RV in the Las .
ose {o,u24 *
Random _ Variable
oN
Discrete Rv Continvovs Ru
— The suppor set is — The support set is
Finite /countabte. uncountable
Remank For ony RV, Se = {ee . Plxe Ce £,2+Ob 0} Lecture 7 (17.1.25)
— Def": Discrete ev
a discrete RV if a countcble set 5. SR st’ R CKES)=1
—~ Probebility Moss _fincton
ler x be @ discrete RV anh FR be the induceot prob measure, ne.
(R. BCR), RB) is the induces pwd. space ank Sx =< IR be a countable such
Ahot Pres = 1, See [email protected]¥
Define p: Sx ~BD st pe = R= = RM Cetxd)
pf =a
© 5 otherwise
Then the finckon p iS cotledt the probability moss finchon of the Rv
~ Bernoulli BV:
Re (x20) = [icp ; if x20 i pe Com
po: iFmeat
©; otherwise
U gy: fs if Fwit ie goest
© ; otherwise
S Bg - exer = ["pterpys Fron.) pec
. ° > othenwis®
=> Carb = = Sex 0% % => Use te prove above ex.
Fa
oa %
Eg- Define pmbobility mass ‘fincton of an Rv x,
Raa =
©; othenvise
Lemma : Foy a. diserete Rv, the distribution foncion is a step fonction
Prof: Hi
— Lebesgue measure.
@ Of R, Lebesgue mersive is the Length
@ on R, lWehesgue Measure is the aren
@ on Rr. Lebegue meogyre 9s the volume
Remark 0 On ®. the Lebesgue measure fF any singleton set is zor
Eg Leb. measme (int) = © Y xER”
@ Oem = 0, @ convention of measyvre th
© Fo% any countable ser ES Rm”, feb. measne(E)~o
fg- Leb. meoswe CU (3) = Za teb- measure ({*9) = ©
© Lev: measure oF ani tine on Qt is ze¥0
— Confinvoys Rv.
Aa ev. & iS soid to be & continuous Rv. iF it fs not discrete
—This_amourts 40 saying thot for ony countable set Sa R, B(cesd
