Scribd
Upload
32 views198 pages

Statistics - I Notes

In this document you can find statistics study material for Masters of Operational Research from Delhi University

Uploaded by

2147 Mahesh Kumar
Copyright
© © 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
32 views198 pages

Statistics - I Notes

In this document you can find statistics study material for Masters of Operational Research from Delhi University

Uploaded by

2147 Mahesh Kumar
Copyright
© © 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
You are on page 1/ 198
—S RaksHur VARMA > _ HINDU COLLEGE ~ —, SECTION A me) —~ SEMESTER ~T 7 —s Stansrics ~T 5 } 4 =. x | WV STATISTI BS ao ll 2 sow: sramisnes -L oh . ap OP I SEMESTER SYLLABY: 7 ‘Peet: Probabidily : Pbobi.ldy Dyiems . Conclit ional Roba ard. Bayes ‘Theerem - : . Random Variables ond their Probablbly Dismbution - ~~“ ™“-“*vVvuVvUUUUUUUOOOKHbEbKG Z : w Characteristic. Function - Hultiainendenal Ronde voriable . Toit, Harve and. Conditieral Distibutions , Thelupendart Random Variables, Fanckions 8, Cweral Randem Variables , Order Stabistics - Diseete ord Contingnws Dish butions - Wenk Law 6 Large Mumbers- Central Limit Thetrem - Concepis 6, Ranelom Sampuiiy: Sample Charasten'stics + Exact Sompuuny Distribubions : Chi-sqpare , t , F ‘dlistibutions - , Parumebie Point Estimation Testing &) Hypotheses and Trtewal Estimation : oblemn 3 Point Estimahien , Hoxium Likelihood Estimates - Simple aret Composite potheses , Neymann- Pearson Lemma - Likedhood Ratio teats. ChinSqpore teats , t-teats - F-tests, Large Sornple ‘teats » Censtrachion. o}, Confidan te Intervals: Scanned by CamScanner a — Ai. i = -_ yO — 25° Ae ED \ se Statistics uh the Science a making decisions a Tre face Dp ureertainty ard-thia, uncertainty panied usin probability. sk Approaches. to ity ore clawixted o8 ! T Catsical Poteet i E- Statistical (Emplrienl ) TE txiowatic . pwromes are known wo * Random Experiment : Experiment’ pthase e cone bat ch ular eukteme tums up ie known Only i ormed. “The experiment wo pejomed ‘we teen a cieeal ensvinmertal cnditon & Tal = Pedonnance 3 a random experiment tb referred to od atnal - + Exhaustive nurbera} canes : Total number ® all thi pasibte eidtemes © a roselom expenmerit - : & Evert : An Event tk the aun wilh hich the random experiment Pesprned 9% Favoumble number 9 cases : Number 3, cases cshith entail the > Ocaurence. & the event under considaration - > 2 Equally likely events : Events are Saicl to be equally likely vdhen > here. & po reason to preypr one to another . PAM exclusive events | Disjoint events : When the. aan evertt “predudes the Occurence 8} arother event + we dwo overts ose reutuolly exclusive - ” sh dependent events : Ty te Ocourenee d one eet doer aot " Bryuance the ocourente, & another evert ; toe Say that the sto ewerts we independent 7 VOUGUbFTKbEEEEDELESESE Scanned by CamScanner % CLASSICAL APPROACH "TD PROBABILITY - 4 a. random experiment , results in ‘n’ exhowshive » exuusle and eq likely ayes Out O which fm’ javeur “the owurence Of, event A’, then probabil | occurence OL 'A' is given by Pla] = i min \ o This approach falls when total number d, cases -tends 10 ‘ mit or When ‘the everts are not equally likely - ‘ ea STATISTICAL / EMPIRICAL APPROACH YO PROBABILITY Ty « random experiment & perfprmed under hemnoyenebuus ard. ‘identical environmental tonditions indefinitely large no. times say, 'n', ed 6 Whidh ‘m’ trials favour te Occurence 4 event $A! then probabil B, otureme 0 ‘A’ th given by PCA) =m , uh a indekndely lary 2 ) 1 ere NN & Ini efnctoly lorry e oR PCA) - tim, a Th other words probability t& the long run relate, frequency (mir) or probability 8} an evett can be approximated by Tt relate. : & Th this approath “the [im ee doe eat Shanty a. Matremation explanation ond Te experiment has 40 be pergorned, ify lage abe Hin Scanned by CamScanner 2k AXIOMATIC APPROACH “To PROGABILITY . \ Defiition was propounded by Russian Stolistidon * dn: Komagerey i Te wes set theoretical operations to +o out probabislif | PROBABILITY THERY -” a Sample 5 + “Totaly a out { Se prssbe oukores 3 a. tandem expehment, denoted by P| T+ can be : Discrete axl Ov conbireus } ts Dne uhoge Clements can be pur wto SET THEORY %& Universal Set GUdhsbbb ser = * A Discrete sample Space S one-to-one corres re volt o sot Hh positive integers - Sean be inde or countably vyaule » ! 2 4A canfinou Sample soue U one hae clemends can foke all pasile | = Values uotthin a given ieterva- cF => % Subset se Event . Tt aan be elementary = or compound. eer - = 2 Elementuny event cannot be decomposed er =3 * Compeurel ovents ore tose than can be : S + Vnien 3 Set & Aug + Occurence J atleast te a, > Aus the two events Aor B. = n > Th General: HUB U--Ute= DAI = Otsurence J afleast one Of => the n ewes Bee Ae ; 3, eves are rahul ma + fw: we Alem tre Hist 2 ar 0 ea > Y Titosedtion 3 Sts Siher ene mn events > * ANB = Simubtaenesus Ouurene 3}, Ane — ALB. =far. mE AZB} 3 Th General: ANAND Ay = fai =f Sirultee ocurene Of Al, ist my " = { ow: oF € AD tele nt weuuJTU su Scanned by CamScanner %* Goplemert 64 a set a AC» Non- ocautence db aC event A- -forrax dA} * Null set: * tb = Frpasible event $ |* sevens DS 4-4-G = Oteurente a ‘A’ but AOB = ANB non-ocurene B’ . | = PBS > fo:ced bu we Bh | + Subset 3} Another-set - | Acs * ACB: Ouurente 3 evant ‘A’ implia ouurence B ‘B'- > CLASS OF Evens SETS EVENTS) Feld ad v- Field - ak Feld: A fietl d a non-empty claws ose which o cased under Bu formation 3} ule union ard complementation. Ty isdenoted by F Symbotically aAev ot BEF then ABE Ah, 8 6 F. oy PLILILIL ISIS IS IIIS SS. Th gine: Sy we F then OH € F od Me Finan This concep ix pr {irite—ferbe—2 Obberele finite ‘sample spate - « ANF € F Ae FE BE * then by dqiition NeFRBEeH then AtuBte & & (AUB). € yg > AnBe F . p-p-anece F: VA BE F tn ASB ec F » MuBE GF > CRURY Ee F > ANB eg F ~ Scanned by CamScanner “Aue ef. sing ACP > MEF =~ en ene Aur ce OS ane p cys om ». aune F Woh 6 then by auiion Ace Fa AOR i | we (none at ay anace F | -— w_ Tes 4 De fell: Non-etig omply don Ss which w dased urder FS te fomation a, countably infuite unien ord complementation A i interval, os * Bol O- field: Sth a o-feld generated. by semi- closed. nterv es Cav) Tramntans Intervals @ alt binds tidudirg single - pat 3 Sd. For oxample ! rs . es Alaa . 2 (-0,a] —. foes (ete) (a,+#) = (-,0] ey (Orb) = (a,b I- fb9 cry Carb] + (ab) + fay a5 *lab) = Cab] + fay-fb} = Sofundion + -A -fundion ushoge cama A eB ove range o> teal cumbers * : w=5 * Piao in P(-T: Sd funtion dhae clemaun 4 A and. om range +1. Cm2 2 Probably sbipet : (0. A, PI) ,also callel Probablity epaue - = Co > 4 AXIOMATIC APPROACH - Given fhe Probobiliy Space (SL, A ,PLI) wage, PLI > & a probabuliy junction if iE safiction [” . AyioM of NON-Neganvicy : PLAV2O WAER T. Ayo oF CERTAINTY : Projet JE. AHOM OF coUNTAGIE aDpTiviTy : Ty ALA, --- Ae om mutually cxdusie ewes mR, tren PL ZA ]> Z Pw) Scanned by CamScanner = P (Bither tre a cuhaly exe everts ) ‘dis 2 Sum total peer indavi dual pribob tes « sk Axon OF Fwite AMDTIVITY * Portiuular core countable addituirty - 5* SEP, wed : oe - tield BA, ox set unthom * Given the sonple spate SL , 00 andl oe jon, the above Ptd.8 a. pablo Hb ¢ axim Of we Prion f cowrtoble ouidibuty 1B also teferecl 0 Axiot OF = ADDITWITY - —_— & Rationay 6) Avions ° Logic. & Polenale behind the axiém Ne in tre Sein o nit a probabiiy auordyy to uthith PI & Leng run relahve freqpendy Thus, ail the properties possessed Lee yoy go on te ne by the nl aon event: » fio OF NON ~ NeqaTUITy vip 9 eR om “p’ os On times wh 9 trials a, te sment uncer consteleration: They, Wa Wy relahwe Tordem reqyoy A peurente of A - New MAYO Sine pide - Ba lony rebhirreg vine There me » Pla] 70. Here we hme The axiom oy Nen- Aegedunlat . al C empirical) SE ee amin Scanned by CamScanner aia }. } ‘9 } eee ee sesso eee eee celle TT) AMON OF CERTAINTY let A- DEA Rela frequen aourene An TM Sime probability a. boy run relakwe {requenty / 1 therefore PCM)= 4 ; Heme we hat Axiom a} Certcurty « TD) Axor oF couwTABLE ADDTVITY A wih their’ respective Ut ALA, --Aybe cajsint everds uv are "' owurentes “As Nan, Mae oa ae random ‘experimert- ‘Then relahie -fequinsy S] Cuurentes % Fa Press Ore respechwely Nae | Ma «Dan nin’ on So ke fying oun’ eo Br etter Nh « a+ Mn =. fen —® ara thed probabi. min ig. a ms PL £x]- 2a) On generaisat on 46 Courdably wfinde number a, events - & Agsgnmert ropabiiti ts glue S- T) when 0. ib dh is chicrete -fuille. sample cpase - DL = £65,082 ---- On} A = Power Gh B Sne 2" Hee we assign nen-neqaiwe quantities Pi (>0) , j=n2 tof Osid rie y2--m such that PID)- PI {4a oat E Plwi)=4 Let A= f Wipe UKE ed Peal= PL § io--- KZ] = SPL] = * Pir eat Scanned by CamScanner henewer te sonple spate & dikuele, -{rule& eniprobable serple-spate axiomatic. approath reduces Yo clawsical approach to pitbabildy D) when Sangle sprue i dbisrde & crunhabhy iii [our Le fon we--% A= Cow 3 all the passible subses 3 1 (Sruhog and 2) ; Tnahis cose Tu not possible to wstqn eqpsl probabidsty 40 Si os Aylom a, certairty would then be vitloked SO here | Ron-neqakive quantities bj zisyaress tb orsyeped 10 ALi; wud’ | Such that Axiom dD, cetaity us salisfiedt » je pit pt: =i Let A be a eomppuncl event aed tren PEA] > BS Plum] fir ien} | Genre ds 0 = feo arma, THE f P Gene PDH] Pini) -PorT] = D A= f Getty af least one heasl 3 > f HH/TH, HTY then PLA] ~ PLHK,TH/HT} = P CHA) + PTH) + PC = ttl+lo2 4H Y q LAPIS IPI FTP ISS IAAT ALS. HT) a twiie Ukely to occur as task gemne 2: Qa ft} Head | Pine, P(m=0 thr PCH)=2 , PCTL 3 3 panes » O° f 4, TH, TH, TITH-+--- } efonh--- f PComde g a $4: Lenk pias Bdedthe os a4 Scanned by CamScanner Ls TE) when sanple space tankinous ° we “iil Ld wee RE lPeseteleg Beane: 0 = fe, 704 P(A) = Tem dn> 1 0 t0 ato P(L= fet de TC Le lo, 10)+ Pons fede rive I > frop= P(r) ye" dn: O 4e PROPERTIES OF PROBABILITY FUNCTION id (9,A,P) bevthe given probabil pld]= PlQ}+Pldl+---- » P[Ol=0 Teoren 2: Robabiliy qundiion ce firilely aabitive - Th other words 4 Aus Py - Ak ore feuchualy excsive events tr A, then fay. Spin) ‘PROOF ; la Me A, lous. be excluswe events § ty A suth thot fied rie Rt; Ke then PT zai] = = pla (By xem 3) courtable at uty) - PL Ev +E o]- Elite 2 rial] 2 P [gai] - Src) Scanned by CamScanner CoRowrey ; 4 Bi, Pe-- - ve exhaustne ancl mutually ercluswe eves na Shen =P [ri] - 1 \ Pe pPPPPPfiIIIIIIIIIIIIISAS. aveselato ey WM ten eos ae estate ont rahully rue ren thay” {orm the tin a Sample PRook: <> DiS are exhaubtue $+ pi “> Bi cxchy we events aA ) > ZPi- Cos Pig ore ruteally , C q courlnble addifunty ord RH ~fatious pre tt oxiem J wy axiem o} Bi weoReH3: PCA‘)=I- PLA} whee dee bea emor, REA , . ACE A = Here PCA) and PCAN) ane a defined! New ALAS = ; » Arpt= D Cs Rard AS are mutally exdiwe } => pCatnt)= P(-%) on pla)+Ptac)= 4 CLs {subs aselom fprde acldite J on peat) = 1-PCA) avienawe: Sine PL ovel AS are exhaushwe rol mutually extlubnt henwe they pm the. partition 3 Saswple Spoue - Hence PCA) + PLAY) = | > PCAt) =(- PCA) ore PRRPPLPL Scanned by CamScanner 3 “Ek bl Say -& THEREH 6 : Probabi Hite THERE A + PRoor : STHEOREN & For any A ea, 0 PCA4PCAS)= | PrAat)= I-P(A) xO 2 \-P(A)%O > Praye | Prdbeitity function us _ moro A onl B are eves nm A then PCA) & PCB) Prof. g-= AUCBAS) > P Sine z | > P(s)= Pta)+ P(BAY) > P(B)= PCA) + P(B-A) > Btey Pla-A)= Ple)- PCA) + Law OF ADDITION OF PROBABILITIES oo > A+ BAS (B)= P(A)+ PCBAS) PCBAt) yO thereprre DY Aan B we events then P(B-A) = P(B)-PLA) B= Av CBR‘) = A+BAS fr Dou Be A, + PCAYB) = AUB + Pca)+ PCG) - Plas) A + @A‘) P-LAue)—p-tay +P tend) AuB = A+B8-AB Pca) < PCB) jon a Subhraohu * th A uth thet #cB, of PIMe | a hot AcB: f DB, B= pa + BAS BAS= B-AB Peay. = Scanned by CamScanner Pauw) = Pca) + PCB) ~ PCAnB) Grenerligatjon : Let At, Pi <=) An be 'n’ ewenks in a then P[Oei]= Pra] -£2 Plaia,]+zez P [ni Bis Bia] lti,singn Isisiseigen meee $f CDEP (AL Pr: An) = SP (ad- Se PLA] + Ss =P [rir Bis Ala] oe list iasivet ket buinigeiser none 4 CPP (APs -- An) Txne2, (Gail - E AW) -SE PLP] Tet jg vheet = pCa) + Peba)} — PCW AL) « 3 ford: pln. 2 Pln]- 22 Plame] + EET PLA Aa As] fet isiwigebet = f Peng +Pena)+ PLAs)} — £ PCAiAa) + Petia) #PUOaPs) +P [AiR As } 2 Bootes SNEQUALITY 7 This in — give, upper bound tothe probabblity %, ouurente y ot least one the “nr! evens - Ld Pi,Pa se An be eents wh A ,then me) eg SP TR: PL Oa] < Zr] ® ‘ ton_th_ Sul addifwe - “De impkes tht Probab lon_ts_Silo adauiure * implies sees ten HC coer Oe rutally exdualve Le we eA AHHHHOTHHEEHEKTSETAARARR Scanned by CamScanner Tor ne 2 P [Aum y = PraltPlm] - Plame) ~ PLA] 7O ~ PERum] ¢ Pcay+ PCa) oe PT On] < 2 PCA Ts © i tre for n22 Let's ayume that ® i the for n= ws Pl bale 3P (m1) —y Assurption We shall new prove that ® & te {Br ne kt! Consider , 7 ‘ PL UA ]- PL YAU Am] F then by axumption 2 PCDI )+ PCAwe) ce PL Om] < = pew) tHoxe ® a deme Sra for n=kt) also Hence by Principle 2 Mathematical Troluch'om Ur vy trae tt . # RONFERRONIS TNEQUALITY Let Pi, Pr An be events wi then Bonjurtmis equality States that Pla] » ZP La] ~ ce) wr: Bi = (GAIS)S By De Nogais law PC Hei) =PC Gat]®= I~ Pisa] 7 TRIS (Oy aig Sie pai te AE AS ,... Ag®) VovouTsHTHHE PIKE ALLBELLLLLLLL enn enncnnemmmm nena, < P( Gri) +P (Pen) Cs © i trefor nod) | Scanned by CamScanner oe = P( AB) %1-Ecr-Peiy = lf k-E PAT » PCAN) » EPcA- Lev Bonjpronis equality quis lower beune| {pr simuttaenesus Ocurens Q vats . lo SEP, Hon eer 2 CONDITIONAL PROBABILITY Seep iionS Let (0, A, P) be the probobilll spate. - let AZ BE A , / Definition : “The conditional ‘probability oy event A guxn that everit B tos ocemrect w& depinecl by Peaisl- PeAd 4 Pwe)»0 ——O Pce) and & undefined! uy PCB)=0 - Hente (> PP CABI~ PCB) PCALB) = Pcp): PCBIA) - AP ion: & Condidional probabil -unckion- . | Gan be gun Relakive frequency Uverpralation « . Le ma, Me ord Mee be number & time A,B and ANG cowie sn’ trials on tandem experimert , then . Ma , MWg & Mee ole reapedindy relate ffeqpandes a} n nm nr fA’ , 7B! ont “AB’ 44 both Pca) £ PCB) >O Scanned by CamScanner can le ne SOHPOHNNVTHNG KH NHKTTKKKL — — I, bhbbbbbi dh 4 4 EEE OEE EEE YG Now Mae tb relodwe freqparcy 9 ocurence ‘A’ the Na hws to which & hos ocenreal ther. lim Megin . PlAB) - PCA|e) ™° "Tein PCB) tlene PCAB) is the ley run OR Conditional Probability dp oceurente A Bearple: Consider oo random cudume thect eowh “Gown eqpolly dbely » conolioval, relabive $equenty on event con be approximated by th onal relative pony: eniment Of tossihy D imps and . Fed - o S G) The probabuty a, 2 heals queen heal on 4ist ¢ Wy ‘The Pres 8 2 hends quien ot least one heoul - + PAW AT, TH TT PCHH) = PCHT )< PCTHS Pom) d Let A : Event 8} gething head on The first coin - onl Ar : Event 8) getithg heod en the Seamed coin - W PCApeth) > POAAA) = PUP) - PLAY) PCA:) fttemowe Qa, = fHH,HT} - PRA. IAi) = + (i) Plame tava} = PCPA (R)] = PUrt) = V4 PCAUAD Atemt: Onyp, = f HH, WT THY <. PU Pie | um | = 4 Wa yy > I Priva.) 34 3 Scanned by CamScanner ATHEOREH : For a quien event Be ef far whith P(e) >0 ‘ Gondienel Probability Function P[-1@] behaves dike an ‘ Ordunory PrOebity unetion. ‘ PRook : We will show thot PL+18) satisfies all the “three. axioms 4 underlay oreuhosy Probobidiy pndtion Ply. 4 5 4 + Axiom 3 dam-negative, ‘ fr AEA ‘ Pirie) = PCAC) (Gy depibion % sardtionel probobsbi } ‘ Pie) ‘ 4a 4 ‘ ‘ « %o . ‘ ‘| vdd« Gime umnerater art denemuiater both sakabyy oxiorm ay, NOP negedry + Prion 8 corny: | Consider = PC-218)= Pras) - PB) -4 Pis) PB) = poxtom 3, (Guntoble_addifualy - be mructonlly exclusive (dhs out) everls Lek Ay Pay (onside, br Eile] - PLCEMB] =P] Zen) P(e) Pie) - = Pine) - & plies] * Pte) ™ Thuy PLUG) abo Sohitfies Ovorn countable auld Scanned by CamScanner ~~ > 3 e~ + Ple]-0 — > P tqisJ-0 . ; k A a ne he zpile)> ZPLAIB FS AN ly > PUAN: 2 | FS hen PE eA] EPC) if Pe. Th NB, ASB => PHB) < PCArIB) RSs PCA) & PLA) m9 Ah ACA — > Ptho-AilB)= P(21B) - PCAIB) | FSS Pp [h2-Ad) = P(N2)- PCA) mS mo k Multiplicadwe low 0 Probabi.ily SCR pereuis to qinckig, Prbubilty & sirulfenecus ocerrence J, =s 3 — P [eits--Me]= PCa) PCAs th) PCAsI AA] ~~ ~~. = voor PERCEBLAL AL Aen] | =e P Epa. Aer] 90 4 i => Sine A Pes CAA, CA oul | q h: ° PCAArPs) 4 PCAAL) < PCA) OR PCA PCAAL) > Pam As) > 0 ——™ => 4 Themen G Teal probabiuty D E,R-- Ee are events im A formuny Partition % the 3 Sample Space ard ‘A’ y Ory event un A then Pcr)- PLE] - PLAS] Der- a & Gigg= > Hileiqaek Now A An & . . -ac veiy= ACeE)= Zhe = = = = ud BVUUUEIIU & Scanned by CamScanner

You might also like