Dr, RIAYKUMAR AR. MODULE -3 - STATISTICAL METHODS AND CURVE FITTING Pearson's coefficient correlation and Tines of regression, problems. fitting the curves of the forms Statistical Methods: Correlation and Regression-Karl rank correlation, problems. Regression analysis, Curve Filling: Curve fitting by the method of least squares, ¥ +b, y = ax? and y = ax? + bx +e. Method of Solving problems on Statistical Methods ‘Method/In Questions Formula | Working Procedure ae 16 7 Find Comelation oF pete — | pit findmean ofeandy. || * 12 Fee] 5=9|% 1" 12 Comet Govasen ve] "JOKER | Const the prope able Karl Pearson's Coefficient ‘Substitutes the XY, EX* tas normal) ae le ee) Cortlation Identify the Byy and By i ‘When two regression r= tyDoPrx the two given Equations and Mean ofxandy~ | sobstimiein r= +YPabox wh No Table lines are given. Two Find mcan of x and y Find the correlation coefficient Take xandy sf and Tinthe sverige | Take xandy as €and J in cree ane or | tegiven two equations and Tandy see Fit find mean of x and y Regression line yon x | Construct the proper Table, yaBEx Substitutes the XY, X= Ee and ¥? in formula YandX | x obwintetneot | Reo ney ons | Sie wo : es Cprotnccrston |= EY trot es pee Seiten onde in i the two obrained regression Fines and substitute in r= tbody xlylaes—y | ay tfiey sven ak et rasksonetn | DG uincale ie ‘and find Sd? , then Sistine oma rhe are not gen ak celica naesendng | order, perform a). Rank correlation coefficient a R, |y | Ry @ scanned with OKEN ScannerDr. AIAYKUMAR AR. Method of Solving problems on Curve fitting Method/Question | Formula/Normal equations —_[ Table dys ayx enh Straight line ae aSx? + byx tee oe Dixy = ape + by x* Dy = abe + BEE tne Second degree | 2AY = @BA* t UAT + ce Parabola Batya ake +80 + ee y=axt+brte| Or = §] dy = na+ yrs cx Ya at bet ce eae? ty = ax? + bY +6 “Taking fog on both side we get togey = loge + bloger =log.a yrax? BY =na+bEx SAY = aDX + UX? Working Procedure for Problems ‘+ We first write the normal equations appropriate to the curve of best fit. © We prepare the relevant table and find the value of all ‘the summations present in the normal equations. ‘© We substitute these value to arrive at a system of equations in the unknowt parameters. © We find the parameters by solving the system of equations and substitute in thy equation y = f(x) which being the curve of best ft. @ scanned with OKEN ScannerDr. AJAYKUMAR AR STATISTICAL METHODS ection, classification, and analysis of onable decisions. It has tal data, etc. Statistics deals with the methods for coll numerical data for drawing valid conclusions and making reas meaningful applications in production engineering, in the analysis of experimen Measures of Central Tendeney and Dispersion ‘Measures of Central Tendency ‘The collection of data constitues the starting point of any statistica’ investigation. fone can be made about the data by analyzing and interpreting its A frequency ‘ofthe data around some central value. Finding this 1e most representative value of the Dei distribution in general shows clustering central value or the average is of importance as it gives th whole group. Different methods give different averages which are known a the Measures of Central Tendency, A measure of Central Tendency gives us a rough idea about where the data points are centered. Mean, Median, Mode, Geometrical mean and Harmonic mean are the commonly used measures of central tendency. Mean (Arithmetic Mean) xy psx, %y ate the set ofr values of a variate, then the arithmetic mean (or simply mean) is given by Ry txptxa tito n ‘Measures of Dispersion “Two distributions may have the same mean and same total frequency, yet they may differ in the extent to which the individual values may be spread about the average. The magnitude Of such a variation is called dispersion. Dispersion is the degree of scattering or variatior of the variable about the central value. Range, Quartile deviation, Mean deviation anc Standard deviation are the important measures of dispersion. ‘Standard Deviation (S.D.) ‘The most important and the most powerful measure of dispersion is the standarv deviation (S.D.). It is generally denoted by a. It is computed as the square root of the meat of the squares of the differences of the variate values from their mean, ‘Thus, the Standard Deviation (S.D.) is given by the formula. @ scanned with OKEN ScannerDd 2) Dr. AIAYKUMAR AR. wher is i here N is the total frequency ¥ fi Note: Th he square of the standard deviation is known as the Varianes es, ice., when the changes in one CORRELATION mndent variabl is called correlation. Such a data . a simultaneous variation of two indeper ee are associated or followed by changes inthe other, i ecting two variables is called bivariate population. Examples: Marks of individuals in two su Height and weight of individuals Rainfall and yield of crops Tf'an increase (or decrease) in the decrease) in the other, the correlation eight and weight, rainfall and yield of crops are in the values of one variable corresponds to a deer the correlation is said to be negative ‘odity and the decreas ibjects alues of one variable corresponds £0 22 increase (oF is said to be positive. positively correlated. cease (OF Example: H Ian increase (oF decrease) increase) in the other, he price of a comm f¢ in its demand leads t0 they are said to be 4 between the variables, Example: Increase in 4 negative correlation Note: If there is 0 relationship indicate jndependent oF uncorrelated. Marks of individuals in 60 subjects are uncorrelated exampl relation (known FICIENT O) cal measure of col ficient of corre J CORRELATION the coefficient of ¢0! lation is calle fined by the relation dis de! rel COE! Jation) an @ scanned with OKEN Scannerbr, AIAYKUMAR AR deviation from the mean &, deviation from the mean F, 2Y and n is the number of values o [1-9 — standard deviation of x, standard deviation of y the two variables. Note: 1) The coefficient of correlation r numerically That is, -1 <7 <1. does not exceed unity 2) Ifr = +1, then the correlation is perfect and positive. Ifr = =1, then the correlation is perfect and negative. Ifr =O, then there is no correlation between the variables. 3) The alternative formula for the correlation coefficient is n(ixy)-(ENEy) (ne*)-29?)x[ner4)-er"] Regression Regression is an estimation of one independent variable in terms of the other. If x & y are correlated, the best fitting straight line in the least square sense gives reasonably a good. relation between x & y. The line of best fit the form y = a + bx (x being the independent variable) is called the regression line of on x and is given by (7-9) = Dyx(x — that is, ¥ x Where X =x-3,¥=y-y,2=2% y= 22 and by, = =“ = regressi icie yx = 57 = Fegression coefficient of y on x. @ scanned with OKEN ScannerDr. AIAYKUMAR AR O-9) =H) bye = The line of best fit of the form x = a + by (y being the independent variable) is called the regression line of x on y and is given by . per (&—%) = byy(y J), thats, X = Fz yeali gilt Where X =x-3,¥=y-J,¥= y= 2 and Dyy > regression coefficient of x on y. (- 3) = 27), Dye = FS Note: 1) The correlation coefficient r is the geometric mean of the regression coefficients. That is, r= +/ByyByy The sign of r will be positive or negative according as the regression coefficients are both positive or both negative. 2) The point (@, 3) lies on both the regression lines. 3) The slope of the regression line of y on x is by and the slope of the regression line of x on Vis byy. RANK OF CORRELATION ‘Spearman’s rank correlation ‘The Spearman’s correlation coefficient is defined as the Pearson correlation coefficient between the rank variables, and it will be denoted by the symbol p (row). Suppose ‘x’ and ‘y’ are two variables, let Ry and R,, are the ranks of variable x and y respectively, and d; = Ry, ~ Ry, be the difference of the ranks then Spearman rank correlation is given by @ scanned with OKEN ScannerFromiems on Coereiatront pe rerenT Fad the Covrelahon coefgrcted ee fee foloet lew, BC 1 3 a Here, nes. a] 3 s| 3 H. TE, wat stnrs, ig > [223] Ss Ys ZY. atest 3e8tt_ ar oy fyes j= EY. 3 ao fies) We Conshraud ates. folldcing dable. xe | yr xN W 4 6 ! C o 3 13 } “2 ° 4 o 4 18 ' 3 ! 9 3 Sot a a H 4 4 ar 7 Ixts [Lys |rxys | . tO | a6 Lis ey Ws ey VERS ce Gey ca) 13 260 the forrelehon co-efficient a gen by (3 (62H ST = 0.80 92 sq nq se 33 a1 1° 6% 3a gy qq 68 8S SH 32 Nea SiGe sco ols kkse n=to zx, 750 [z- as) - ZY. An 2) n o . le Conphod tte felloory fab 7 Xe xz] Y a AS B 86 It 1) asqy | 176 83 14 u “196 yal a. 12 19 iy | él +4 u s 1al as a 68 & 4 ou 16 “38 as 2 13 4H 164 se 7s 18 as | aan | BA “12 ip ra | wee | 22 3F “AR “35 he | i985] a0 oF . ras as bas gas | 2as | PERE Tyee | LAY = BOT6 | RF6R | 1FB The cpvreldhion co-efft cxew? 2 qe by wpe EY VEEN) 73H _ 184 = o, ee = “guoeor ~ CoS” \ 0b x 2962 Thug fhe @urelalon Coofficxe <0-FHIS. @ scanned with OKEN ScannerDye hologped dele oy paletirpence aud 9) ergs ae were eupp Heed’ do to shack on to felto5 a 4 showr dtrod dea indellrged vce ax wedto. Coleedie dre co“ g(ccred ay cores HO 5 Tela} ato etete le le & | s 7 Tal wl It tA Jd er) spam toy | to2 | fol (03 | (00| 48 2 tO | os f -t | eas] ol oS | = os ' 4: y LS O° = ae mi Le | The oe = Er +089 S fadcten coe Dives Nas 0:60) Tht qtr Cove @ scanned with OKEN Scannerfor dee ello Nshion ¢ S Find fee Coefficrewd of conte deDet as fa fen f fas ead ab = : 26l = ws Ig lo | eA | 6 30 | z| Avs SF 0-60 pera ay (ees 8S) tind tte Cooreldhon coe fpecred L/w fr dee folfocer ~ : | fr 1 “S r 613 | q_| af my Pe] t 3 | 5 ta] \ ate fete a ly | ites dng, 201} ) ay pron D Find the Covseloion coe ffrere blw qr ri [oes af [y_[ias | 138 dts presi dla F: Safi ; er: Nervous berg wy ) Ten peop® 4 : ue Pipe eset hoe lelfer# Boks a as Yard dizdanice The number a tebber Corre ty wcod a below. gt in perf sl mae Ss a eeh roles [eae pa 6.) Aw: ys -0.08S @ scanned with OKEN Scanner@.m follow rs dalelasig' te to sdededdg marver © oe Peor zon’ x cowppecre? ry Cove to aget Aye 9 haabowd a3 ar rs 29 30 > j wife 13 2? 23 74 ze 1a to) Finck do coeffere? e export , pro cducton ale Podude? Tee thins) Exons ae) OD 25) 2f| oF pir fats do aqe Colerrtve Corseletion bf aig tO ge SS; AS Gye SB oe. 27 Coco) Ee iar pf dher 2) 3 = 36 os haat aos 24 32 wo ee pcutbed) fowiny dela go 60 60 ng RS nS @ scanned with OKEN ScannerPeoblews on Grex 4 Dd The hve veqre sion wank Y are Ser (BIB O° Find 1) mean of 20'S Ww dte @rreloien coeff’ ai" = We have the, reqre* Be 1UIB —0-BF¥ =O Ye (len 0. so Se 1 ->@, clearly fle veqression bay > - 0-84 aud equation’ er Capron 4 die Varrabl say ame fe 1 64-0.50% and uy mean oy ae cred Cnet (Raqrenrion Cu 4 n bn oe wees raha ton 0 Rogreb we coegprre™ are by oe 70" sO pote fee rege Since toe poh (zp) Ofer One, We rate Qe qos Ubu - 0.se above x 40.8499 21%! Thus meansop 2% mean of WP The Cosreladion Coegoeter? rv aa 52 tl bay by x =o 2 14.13 —0-874 ca 4 equeiem poten FL 3 ver ey @ scanned with OKEN Scannerwea 7 = + 0.66 Smee bode 1s wegren Hien co egrore* are neqhive , we hove we 0-66 Thus, dhe Coscelaion cogected we 066. clenfoyed Laboratory record) wz andy teny renpion 39 ee avd goxt-W* ae SY 4 337° ens coeffecte > nwo portally only die Cres of avorGbl wey Hw fale 2,4 are for, PE ppeciuely. Caleu ef cosrelaren peeee ant # gion Lined ae Zr We howe tre weT™ Hoe- s¥ + 3gso 7 SH yee gox - AY = 104 => gor = ot TY Revndtivy to Giese gan Vis 660+ 080% 20 ae eee vom 2® Cleon Joe wegen ere f pay = Oks ond byx = 0-80: smce tht pow (2,9) lie en bod. dos regre # tier lina, “we have nt =e 22> gox -17 = 07 oboe © ucony , we a and Teel. @ scanned with OKEN ScannerThe correlahion coef{rereud a aver boy 622 Sbay byx = t J 6us) Co) Y= + 0.6 Since boda foe regres ston we hove yeab. are oxi Co ef ore wy ploles. sim plefeing fhe Ropexton lin oe dis qe rene B tre bile simpbfyiry ohn take alley loxge co-efferod 4 afy on Ber ecamipl = Refer Goble D . He a Nal -SVWe-3B > sye +33 ae Seared gon e 2D goxrz= ort 20K - IY =lOF eee 825-12 EY ig Hoe e- 33° ST 2 , _ D ye - sT 4 2222 > always lea amd bye are qreser am one C effrcrert pitt be orl o- Seoppoxe Tye -10F 420% becausey op bay ghan dee eq Conrelatre rho rod pos eibls. Alor Zo, te, ~l4rel. “f ocony ax Co efferet QD lorsyacred- Phan M co effrcre” i dee quer tow byx 9 Pogreaon Gis qd Y onx [# y Co-effiecend & larger doar 2 Co-effecred im foo quen tre J boxy 2 Raqre pion equ bibs @ scanned with OKEN ScannerCones 4 foo Mom equations of regression Saget a eee t 43 = Yo psig x + 33-435 | Wz O-Slay+ tnd ft Correlation Coegprererd and dhe meame xX and y, : sol” We haye die reqrex dion Gra Goo.SI6X + 33-43 = 0.SIaY + 3a.SA Cleawly te weqrespion co -efproved# are BaGeroisre enced yarn ors ° : ns the Since, dus poir (#7) been ca veqrex4sien Cun, CWE ala MONE = OSIEXEY =BS.93 Xe — 0.29 = BAS2 ; 5 Sokuvy chal abbeys aqudlow, WE BS ; ee.6s | Gizepea] ws [LE Thur, die mmean”o a avk.Y, The Correlation Coeppretent poet J bay byx = t fo.si2 xo.St6 Be oh qo Ly y at 0. SIH bode fee reqreneten Gefecredx a fave, Ye o-s1y Since posite , we on, tte Gr elon Cepecen = O-StY @ scanned with OKEN Scanner» Follow nay yecordp~, ave oxtailaok. Reqrcwos i SX joy + 66 =0- Moxey 22, SAF 3. Find © mean valeec 4 = nude y) Oaoy ® coegfrered 4 Corr 07 0 We hee dts wegrension Cred Soe - (0Y +66 50 2 l0y 2466 +B Hox -1wy-21y 0 = HOF gin +i8y Sunply fying dee abee equtiet Yr 66 406% He saspousy coin ety gpven epee Atal TEITE D Snce tee pot Pern ood ee “7 Gre, we have oY 266+ O% 7 2 Hox -18y = 24 nox = 2tH +18Y sSelWing ths above equstony, WE yt ze) Dd The, Covreiadion covetivetend 1 DH yer by wed foxy by oF fousxoe = t06 Since bette dix veyrertlon co-effrcseds or% aatt we hone freo-6 I by roy TOY b oe U8Y = — aby > ONE Wet, Sad ao =a Y Y > bz -oY 66 ) Tegud ey Ov O8K3 O-6 oye 4 | @ scanned with OKEN Scanner\J The ; euont aud y ae } eypseccion Ciny too Newrdllel ten ¢ t YS ©. Si6x4 33.93 2 O.SILY + 22182,.6 wervelahon og Gfficred — aurch fe Vitcrmt ey x ” t, ; We be 7 ANE tts weereuron Gnu : SIE X + 32.3% x ° SIZ ¢ 32.52 CHeseby dua trtemeuton, ce eppecrerde an leoyoe= O-s16) and [bey = o.srd Th worrelebros Coeprcet a qrver by lo.st2 xo.Ste ees Ov Si hy Rae V boy byx 5 Since boda due ea e qreuto 5 netewy are Posrks yreuton Co-eff * og ede ve We bere sy Smee tos pow (CS) (ieee ek ee Cfo (OSS Serene Gye we have OsIbe-Y =~ 32.43 % -0.5125 fae SS? 5 Sclumg tr ateve equation , [2-69.63] and We sab @ scanned with OKEN ScannerRX43725, Calculde D mean yatue of x and ¥ q aeV4 ® tes ; FOS Co- efferent a CovwelSion . Avs 2 -0-%6b DD The do vepreuton eyettont ay fer Vawrabler oud Y ave wea yes-2 and Yoox +23 Final > mean ef x audy D correthon Co-efprered > Two random varabl. have te Brrry= 26 and Ext 4 #2! Find toa mean Ve aud fe @-efpcrerd of Covrelateon Ofro HRY, Bu fen, Yo R THOS deBoyed laboratory dele. 2 wegrewon Yen + a x on ¥ are Lvtar la bl or FxX-1bYtTLO aud sy 4-3 FO wecpemruely, Culouldde x, 9 cd Me looper 4 Covrelaon bdveer x AY, Dare comgaty ae et ea lines only ew Aw 2% 20-1034 Y= o-ssar Ts 0:9381IT AO @ scanned with OKEN Scanner3 ? Obketn die Grete) vesrennion aud hence de Tee conretadior Congyrene wd for doe elton dale ect! 12,3 fHis Siefs [3 [sa] 4 so} eh Here, snes. cs) Ze De, oesenes, 1S a l/B@ oe se [Gee GF. YL gAtspsrxwey . 2H 2 2 fist athe We comadract die qeltesing deble SET Rew s f yes-3 | x | Y ze-3 | eyes | + ; 3 i 2 -2 =) f o 2 2 ° 3 ‘i ° a2 o n 3 1 9 GI & \ | 2 y | A a2 7 <4 [exe | Zy'eas | B=" ; suen bs The equolront y vegreastion bna ase aver y ton Line ay Mt ONF vespeseion une “4 Yon mee BE IO 4 =A yea Cy-4) = zxY (2-2) (x)= BLO 9) — 13 Cy-9) G@-s) = 43 4-3) &- 3) = Be u yes = 103 (t-3) mo BF os (4-5) ye Bde BTS me os - as +> Sele Vist ede wee OrS 4oSY sus, Ye PP e326 and oer 0-S40-Sy are Tus Weqy ce ARON Gnex. @ scanned with OKEN ScannerClecaby tus meqrexxion co“yrcreWa ore bag FOS amd bye = 13 : TH Covveladion coegfiered 7 9 Ye 79 wet Jb, byx eatJfosKrs 2% 08062 t Smee post the wre spreanion coopierewla ane po ( Weheve ars 0-806. nd fh D Oblain tea Gwr ef veqrersion aid ds of covreldhion effected fer 7 ge! rep ye ppals[e] 4) (a [st [zs [ol a [3 4] ss Here, O24 - g- te. ao feen | 41" = ng tate, We Conghud dus qplleers 4 a 4 Teel ea | Feasts T | 4 =3 | ass | H | 6 | a 8 abi) 73 4 a | 6 | 13 | to et | -) 1 \o4 | \ | A 12 ° \ ° | \ oO | js iu \ 0 1 | 0 | o | 6 13 2 2 4 | # ojo | 404 3 3 a | Ze=93 |TPsas | rRyras [ _|—_—__——_____ Fla reqreazron Lig are quen b ; were sBIon line You DO G sweqre pate Gre 2 . LRKY ~ 2xY Vielen ie Xe Se y Ce) = ERY Gx) (x-%) = EXY cy. x? xt @ scanned with OKEN Scannerone 2ECG) eu s lore sa Yoo oqgx-3-42 +0 we or tsy AS Yoo Has 40.93% cen bras torts Thuy, Yr Arepodaxe go xe gaz rorsy are te “Peqrenx ren Gre Clearly, dae veqrenaion coeffrcreWa awe lomeg= 0-43 aud byx = 0-93 Th Correldhion coegecred s B giver oy Bet Vbay bye = + [043x093 = £ O13 Since bed das veqrension coeffrereds awe post tive We heave x2 0-93 — anol 0x) dwo hughand > The following doe giver dos are af nee and fire age ef wife Cy) Dm Years. Form dhe weqresgion Unr curd Coleulote dea age of Corre ponding do lb Yearg age 4 wife- x] st [a3 [ar] as] as[ar| 30 [31 | 33] 35° ic ar| 18 [ao |aalar|ar [aq | ar| aq| av zi Here mz10 xy = _ FX _ 300 = fry aa we os Te Ppee30 ay rs Re cConghud dee ellos dabl. @ scanned with OKEN Scanner- 4 | “= [Vewal x2) ye lay | ; 36 ~ easy is | ¥ | 1 aq 4 36 | Ve | aH } Qazi = ny | na [Ra A | g0 ss | | as | ts | as | aa on H | 4 | 6 \ 2s | ay Tn | 7 | al ‘ ro) te | 4 | so | aq rn o | te | 6 | 3) | at i Ho] 2 | 33 | ag a a | a | 3s | as Ss as” A | is \ Lt bang av Be The | equotions ofp teqrenzion bia ore gave oo Th, Beqremwion bis Yons & reqreazron hiv & cad ys ZRY x2 BY y Txt ay Ce-G)= 257. wx-2) cx-#) = EXY Cy -9) LK zYy* @-as) - B3 (x- 30) 30 = 123. (y-as) @ ) 138 mx-30 164 c FE OBA BE FAIS as ovasy - L3-ASH2 Ga Hb ORIN mes the 24 DO FY Ths , Gee sroBW Lo xa thrsto.ay ave due weqTe BRON Gre, pe ned do find fe value 4 x when qe So we moy vse die equetkon pes 124 O-ASY ho fered oc phew Yate. clearly vel Yar6, we obbeun = Ure + 0-980) = 23-a5Y%Q3 Thur ta age 4 Aug band YW 83 Veors rhe dia oge qh, eo Peter ence, @ scanned with OKEN ScannerED Obleam ave Unto veyrex nto and Coc gfectew 4 Comrettien yor ste ellos ade =a) | a eRe} 8 | 10 | 4 ) Om aie | -10 geet [sn mH | 1a | 6 Saale aa | 6 1s | an “2 3 | 4p Tl 0 a2 | 6 2 -s | H | pas” =a ab | 30 6 4 36 | sl 5 30 | 36 10 is -| 100 | aas_| 'S 1 rt Fea rae Ea | The Lwa 4 veqression ore . weqrension Cus of Y ont a “- q~ a Ye oe x x= a ze YY ; Cy-4) 2 ZEL ce-8) cx)> Eee cia) Ext a year = 8S (x-20) (xt- po) = BSR C4 an) ago Ys 09-18 al pleny seuss ye Bt ote ee 6 +OHG Thaap, spe Bt01% He HOLY ane dee reyressl clearl,, tts wegreszion Co- effrcreudy are bxy>OrH The Correledion coe ffected Y ave boy we bey yx atfouxo9 =t0.6 Since bode fle weyzazp on Cocfpittewdx are popie he have oy apie Mi P @ scanned with OKEN Scanner< Find dhe ner a Correlahon ce eggrereut der lil alaly | 1g |] 1a] wl] asl as| 36] dna: 5. GEHVtH BI GL = oT Ab bees] s |6 |#18 weapreneron amd hence find the Cas elloeing, Ade Tate iq| HO| SO -or64 tory & 2 O16 aI] Obleun tue veqrensien Gner aud eco) ee ths. Correledion Coeppreied Jor. dee follow ing ea 3 | HR Ls 5 rato | ts | ele fells fia [te [te [eo | 22] 32 | Avs YQ St btexr BK xe oUtonyy HK F088 t+. find dhe vere x BION equates yonx amd xé vag dee dable of verlues qe below 16 ay | 32 | no [ug [se | 0°39 O-AS | [23 a) | 2.24 | 3-3)| DC i @ scanned with OKEN Scanner