2 g INTRODUCTION -The study of regression has special importance in statistical analysis. We know that the jationship between OVO series is measured with the help of correlation. Under correlation, the. ‘rection and magni de of the re 0 onthe come the best estimate of th iedependent variable by corre in the study of regress timation. 8 lysis is very important and useful. m MEANING AND DEFINITION According to Oxford English Dictionary, the word ‘regression’ means “Stepping back” or “Returning to average value”. The term was first of all used by a famous Biological Scientist in 19th century, Sir Francis Galton relating to a study of hereditary characteristics. He found out an interesting result by making a study of the height of about one thousand fathers and sons. His_ conclusion was that (i) Sons of tall fathers tend to be tall and sons of short fathers tend to be short in height (ii) But mean height of the tall fathers is greater than the mean height of the song, whereas nigan height of the short sons is greater than the mean height of the short fathers. The tendency of the entire mankind to twin back to average ‘height, was termed by Galton ‘Regression towards Mediocricify” anid the line that shows such type of trend was named as ‘Regression Line’. Instatistical analysis, the term ‘Regression’ is taken in wider sense, Regression is the study of the nature of relation between the variables so that one may be able to predict the unknown value of one variable for a known value of another variable. In regression, on™ variable is considered as an independent variable and another variable is taken as dependent me With the help of regression, possible values of the dependent variable are estimated on the asis of th i . values of the independent variable. For example, there exists a functional relationship an neon demand and price, i.e. D=/(P). Here, emand (D)is a dependent variablé, and price (P) is ‘independent variable. On the basis of this rel lationship between demand and price, probable “NSS of demand can be estimated corresponding to the different values of price. ™ DEFINITION OF REGRESSION Some i ise ‘me important definitions of regression are as follows: J Rees Sion is the measure of the average relationship between two or Eat a Regression analysis measures the nature and extent of the relation between two cs more ‘atlables, thus enables us to make predictions. 54 & 8 Business Sty q joni istical method of studying the nature of relatj ; In brief, regressionis a stati Ig lationship beteey variables and to make prediction: m UTILITY OF REGRESSION The study of regression is very useful and important in statistical analysis, which is clear ' hy yy following points: (J) Nature of Relationship: Regression analysis explains the nature of relationship bey ‘ee two variables. (2) Estimation of Relationship: The mutual relationship between two or more variables my, measured easily by regression analysis. (3) Prediction: By regression analysis, the value of a dependent variable can be predicated the basis of the value of an independent variable. For example, if price of a commodity rises, yi will be the probable fall in demand, this can be predicted by regression. | (4) Useful in Economic and Business Research: Regression analysis is very usefial in busine | and economic research. With the help of regression, business and economic policies can te | formulated. @ DIFFERENCE BETWEEN CORRELATION AND REGRESSION hee difference between correlation and regression is as follows: ') Degree and Nature of Relationship: Correlation is a measure of degree of relationship ship between the variables so ta between X and Y whereas regression studies the nature of relati one may be able to predict the value of one variable on the basi: (2¥ Cause and Effect Relationship: Correlation does not always assume cause and effec telationship between two variables. Though two variables may be highly correlated, yet it does not necessary follow that one variable is the cause and another variable is the effect. But regression clearly expresses the cause and effect relationship between two variables. One variable is considered independent in regression, for which the value is given and other variable is dependent, which is estimated. The independent variable is the cause and the dependent variable is effect rediction: Correlation does not help in making prediction whereas regression enable us make prediction. With the help of regression line of Y on X, the probable values of Y can be predicted on the basis of the values of X. (4) Symmetric: In correlation analysis, correlation coefficient (r,,) is the M*4SUTE of direction and degree of linear relationship between the two variables X and Y. ‘xy and 732g > mmetrical, Le. Py =Fyq. This implies that it is immaterial which of X and Y is dependent VOT? symud which is independent, In regression analysis, the regression coefficients by 904 Pos To aression eeceies Les 4,, # b,,. Thus, correlation coefficients ra, and 7}. are symmetric W! ficients by, and by, are not symmetric. wurious OF BO-SNSe egy, (5) Non-sense Correlation: Sometimes, there may exist SP 0n rise in income ang elation between two variables by chance, like the correlation. if Y.) there is nothing Tike ng, Ie in weight is a non-sense correlation but in regression analy nse regression.Lineage ASSN Analysis ‘Bresson: ()-Simple and Multiple Re, vatfables at a time, in which an example of « in multiple regres ind others are j imple regres sion anal: uation, other variable in a changing ratio, then it is referred relationship, expressed on 7 @ graph paper takes the fo ee equation, rm is often.used.and. under this, Regression Lines, S concepts are very importaiit to be studied, lationship between two variables. This is also known as sis of regression line, we can predict the value of'a dependent variable. the given value of the independent variable. If two variables X and Y are given, then egre related to,them which are as follows: ¥: The regression line of X on Y gives the best estimate for the R aie, imate for the Q Y : ion line of Y on X gives the best estimate value a seression Line of Y on_X: The regression lin tY for any given value of X. . Nature-o} ~~ With the known i ression) f Regression Lines (or Relation between Correlation and ae nie help of the direction and magnitude of correlation, the ee *“"© main points regarding the relationship among them are as fol80 Business Stat ie ; if The two regression lines are coincident or there will be only one regression line if, Le., there is perfect correlation. This is clear from the following diagrams: a ‘, areal | rat “ + +1 Poe vo oF (2) The two regression lines intersect cach other at 90° if r= 0. This is clear from the diagran given below ‘ (3) The nearer the regression lines are to each other, the greater will be the degree of correlation. | om the contrary, the greater the distance between the two regression lines, the lesser will be the | Gcgree of correlation. This is clear from the following diagrams: y ris low x & x yonk y x o < Ls (4) If regression lines rise from left to right upward, then correlation is positive. On the other side, if nese line move from right to left, then correlation is negative. (5) The regression lines cut each other at the point of intersection of Xand ¥ . This is clear from the following diagram: —_—Analysis ression sap Re pine BS . @ Methods of Obtaining Regression Lines (1) Scatter Diagram Method, Q) Least Square Method. > (1) Scatter Diagram Method Thisis the simplest method of constructing regression lines, In this method, values of the related \ariables are plotted on a graph. A straight line is drawn passing through the plotted points. The straight line is drawn with freehand. This shape of regression line can be linear or non-linear also. This depends upon the location of plotted points. This method is very rarely used in practice because in this method, the decision of the person who draws the regression lines very much affects the result. > (2) Least Square Method Regression lines are also constructed by least square method. Under this method, a regression line is fitted through different points in such a way that the sum of squares of the deviations of the observed values from the fitted line shall be least. The line drawn by this method is called as the Line of Best Fit. In other words, under this method, the two regression lines, are drawn in such a way that sum of the squared deviations becomes minimum. The regression line of Y on X is so crawn such that vertically, the sum of squared deviations becomes minimum relating to the different points and the regression line on X on Y is so drawn such that horizontally, squared deviations of different points add up to the minimum. & Regression Equations Regression equations are the algebraic formulation of regression lines. Regression equations "Present regression lines. Just as there are two regression lines, similarly there are two regression “Stations, which are as follows: on he Regression Equation of ¥ on X: This equation is used to estimate the probable values of Y «© basis of the given values of X. This equation is expressed in the following way: Y=a+bx Her ® @and b are constants.82 Regression equation of Y on X can also be presented in another way as: ¥-F=r.2" (x_¥) oy or ¥-F=byx(x -X) Hete, byx Regression coefficient of Y on X. i i is equati to estimate the probaby | (2) Regression Equation of X on Y: This equation is ba _ nari st e Va, ty on the of the given values of Y. This equation is express: ry: X =a) +d Here, agand boare constants, Regression equation of X on Y can also be written in another way: X-X=rSe y_py oe or X-X=by(y-7) Here, bry = Regression Coefficient of X on Y, © Regression Coefficients Just as there are Regression coefficient abe change in the value of one variable fora unit chai the value of another variable Regression coefficient. line. For two variables X and Y, the q@ Regression Coefficient of Y on x: This Coefficier ; at shows that with a unit change in the alu of X variable, what will be the Average change inthe value ot y arlable, This is represented by byx. Its formula is as followe bwo regression ¢: s, byx =r So oC, The value of byx can also be determined by other formulae, 2) Regression Coefficient of X on Y: This coef te oS that With & un the average change in the value of X-variable: tp bxy. Its formula is as follows: by =. The value of bxy can also be found out by other formulae, D Properties of Regression Coefficients The main properties ofthe regression coefficients are as follows: (1) Coefficient of correlation is the geometric mea i of the regression OeMicients ie. r= y[bxy x byx. wi) r Regr ssion coefficient of Y on X(byx) = Ali) Multiplying () and (ii) %, 6, by. by =r 8, 6 by and bay have positive sins, then r will alo take plus sign (4) Both the regression coefficients cannot be greater than coefficient of y on xis greater than unity, then the regression voeffe unity. This is because r= lige ay = and never greater than one. If both the regression coeficiens happen tobe more than | then ther fSeometric mean will exceed 1 which will not give the corelion coefficients whose value neve exceeds 1. unity: If one regression ‘ent of xony must be less than (S) Arithmetic mean of two regression coefficients i ether equal to or greater than the correlation coefficient. In terms of the formula: byx they 5 2 (© Shift of origin does not affect regression coefficients but shift in scale does affect regression coefficients, Regression coefficients are independent ofthe change of origin but not of scale. This means if some common factor is taken out from the items of the series, then in that case, ‘We will have to make adjustment in the regression coefficient formula which is shown belows yx =byu.2 and. bay =buv, Where, w=" 4 andy= "=P and i, and i, are common factors of Y and X series respectively.Business Sy 4 ‘ @ To Obtain Regression Equations : atts The computation of regression equations can be divided into two pé (A) Regression Equations in case of Individual Series. (B) Regression Equations in case of Grouped Data. >( (A) Methods to Obtain Regression Equations in case of Individual Series \ " : ethods, whi “In individual series, regression equations can be worked out by two m ich area follows: Regression Equations using Normal Equations), egression Equations using Regression Coefficients. > (l) Regression Equations using Normal Equations ; This method is also called as Least Square Method. Under this method, computation of regression equations is done by solving out two normal equations. This methcd becomes clear by the following: Regression Equation of Y on X Regression Equation of Y on X is expressed as follows: Y=a+bx ‘Where, Y= Dependent variable, X = Independent variable, = Y-intercept, b= Slope of the line, Under east square method, the values of a and bare obtained by using the following two normal equations: LY =Na+bEx EXY =aLY + 55x? Solving these equations, we get the followin ig Value of a and b, Finally, the caleulat ed value ofa and is putin the equation Y= a 4p , of Yand X wll be used to estimate the value of Ywhen he value of x sete "Seression equation Note: solute increase Regression Equation of X on Regression Equation of Yon Yis expressed a follows: X05 +bo¥ Where, X= Dependent variable, Under least square method, normal equations: Y= Independent variable, a,= Aintercept, b= g + b= Sloy the values of ayand b, are obtained by using the oa line, ‘Owing two EX = Nay +b2YLinear Re S10 Analysis, +byry? Solving the Quations, we Bet the following value of ayand by. by=b,, NENY —exzy wt 7 _ M2Y?-arye & 4) =X ~ by L Final, the caleutated Nalue of ap and by are put in the equation X'= ao + bo¥. The regression equation of X on Y will be used to estimate the value of X when the value of ais the X-intereept, which indicates the ofthe line or called regression coeffi The following examples makes the above Note: Y is given, “ay Xintercent ‘minimum value of X for Y = 0 and byis the slope ‘ent of X on Y, ‘aid method more clear: Example. Calculate the Tegression equation of X on Y from the following data by the method of least square: us 1 2 3 4 5 ne 2 5 3 8 7 Solution: Calculation of Regression Equation x x? ¥ y? XY 1 2 4 2 2 4 5 25 10 3 9 3 9 9 4 16 8 64 32 5 25 : 49 35, N=5, EX=15 | yx2=55 BY =25 by? = 151 EXY = 88 Regression Equation of X on Y is X=a+by ‘The two normal equations are: 2X= Na+ b3Y EXY=aEY +b2Y? Substituting the values, we get 15=5a+25b O 88=25a+151b + ii) Itiplying (i) by 5 and subtracting it from (ii) aki 88=25a +151 75 = 25a + 125b 13=26bAliter: Putting the value of b in equation (i) 1S=Sa +25x0.5 15=Sa+12.5 Sa=25 a=0.50 “NX $0540.57 The value of a and b can also be obtained by using the following formula: =NEAY-EXEY sg yy NEY? -(ey)? Substituting the values, we get b= 5X88- (1925) _ 440-375 _ 65 _1 ”“Sx151-(@5)? 755625 130 72 3x5=3-25=05 2 5 *X=05+0s¥ Example2. Obtain the regression equation oFY on X by the least square method for the following data: x: 1 2 3 4 5 i 9 9 10 12 1" Solution: Also estimate the value of Y when X= 10, Calculation of. Regression Equation of y onX | = x xy x 1 9 9 1 2 9 18 4 3 10 30 9 4 12 48 16 5 1 55 25 =X = 15 XY = 160 =x? 55 Regression Equation of YonXis| SSS _| Y=a+bx_ “a2X s5yy? | Substituting the Values, we get Worse ‘a - 1504555 . Multiplying (by 3 and Subtractin, ° 160=15q +55b 153 =1Sa4455 8 it ftom (i) 1 in equation (i) Sl=Sa +15(0. 1)=5a+ 10.5 Sa=40,5 a=81 ‘he required regression equation of y Y=81+0.7x Estimation for Y For X= 10, y= Putting the Value of b Hence, on X is given by 1 +0.7(10) = 15.1 Given the following data N=8,5Y =21,2x? =99,5 Using the values, find (8) Regression equation of Y on X. (ii) Regression equation of X on Y. (ii) Most approximate value of Y for X = 10 (iv) Most approximate value of X for Y = 2.5 (H Regression Equation of Y on X Y=a+bx N.EXY ~EX EY _8x36-(21)(4) ee = 0581 NEX?-(XP —8x99-(a1? Example 3, LEY? =68,0X7 =36 Solution: byx Y=~ 1.025 +0581Business Sey % 88 con (ii) Regression Equation of X on = ay + byY ‘ sty ms? 23-4 0356 by= waY? (ary 8x68 -(4) 7 ay = ¥ - by¥ = 2.625 -(0.386)(0.5) =2. X= 2.432 + 0.3867 (iii) Prediction for Y | When X =10,¥ =-1.025 +0.581(10)=4.785 (iv) Prediction for X When Y =2.5,¥ =2.432+0,386(2.5) =3.397 EXERCISE 2.1 1. Obtain the line of regression of Y on X by least square method for the following data: & 1 2 3 4 5 Y: 2 3 5 4 : Also obtain an estimate of ¥ when = 2. (Ans. ¥=13+0.9%.3 ny} 2. Find the: Tegression of Y on X and X on Y by the least: ‘Square method for the follewing data; x: 1 2 3 lin : — Also find Coefficient of Correlation, 0, 357+0.643Y;r=0.983) — [tea TP 8 Le aay LY @ependent variabioy 1 2 4 5 PT 4. Obtain the two lines of Tegresson from the following data; (88 Y06nr5 0.64) N=3,E¥ =6,5y? =14,2Y =15,yy2 =T1,EXY =3] 5. Given: 2x =15,5y =] 1OEXY = 400,n¥2 Find the following: © Regression coefficient of Yon Xand the Yintercept, (Gi) X-intercept, and the regression Coefficient of X on y. (i) Most approximate value of Yt, X=5, () Most approximate value of X for Y=25, (Ans. () = 1.033, @= 451, (i (Ans. y= 9.54. +4, Xe 05y =250,2Y? =3099 05) =10 i) a= 0.201, b= 0.118 Wi) Y 4 . S616 y =3151]Linear Regression Analy sig 89 Bression Coefficients Y vdyy 9 O™PUted with the help of regression Coefficients, For this, 7 ion cont Bx from the given data, Regression equations can_be i setts by any oFne following methods: Using the actuay Values ofX ang y Series, : (2) Using deviations fron, Actual Mens 8) Using deviations from (4) Using ro , (Ising the Actual Vatue Inthis method, actual OE ade oa c ‘Used to determine ro to regression 275-225 50 Regression Coefficient of X on Y (bxy): y= NeEAY -BX.EY N xy? -(eyy? = SXBB=UING) _ 440-375 _ 65 _ 9g 5x151-(25)? 155-625 130 Regression Equation of ¥ on X Y-Y=byx(X -X) Substituting the values, Y-5 =1.3(X -3) Y-5=13X-39 Y=13X-3.945 Y=13X411 Regression Equation of X on Y X-X=by (¥-Y) X-3=+05(¥-5) X-3=05Y-25 X=05Y+05 x x? Y ¥E aie [ 1 1 2 4 2 2 4 5 bd 10 3 9 3 8 8 4 16 8 64 eee 5 2 7 49 35 N=5, IX=15 2x2555 EY=25 5y?=154 Ex=89 — | EX 1s NS Regression Coefficient of Y on X (byx): | N LEXY -EY.5Y | byr= 5 > N=10 _N-EXY -BX BY EAY/N-X.¥ byx = * N.EX? -(xy’ o 350 70 35-305 4 404 =125 byy=N-SXY-EX.EY | EAY/N-¥.F y= EE gp AE N.zy?-(y)? of - x: = 5 3 2 4 1 7 | 4 1 0 0 sa.[ 2 1 3 ia - . A) Fit a regression line of Yon-X and predict vie) (ii) Fita regression line of X on Y and predict Y=25 (Ans. ¥ ==1 O25 + O81 285784 : 2. Byusing the following data, find the regression & 785; X25 = 3.397] 0. Y when X=10. onXand compute the valueof X 35.5, ¥ =4.0, EX? =385, EY? =192,5(X +Y)? =947 and 3. Given that 250,2Y =300, 6, =5, 6, =10, EXY =7900,V =10 Compute : (i) Two regression coefficients, (ii) Correlation coefficient between X and Y. (iii) Most approximate value of ¥ when X = 55 and X when Y = 49, Tans. dyx= 1.6, bxy= 04, r= 0.8, Yog=75, s0* 29)inear Regression Analysis 95 4. By using the follo 5. Following information y, n Wing dat and estimated vatye ¢, "8, find comelation Coefficient and regression equation of Y on X N=10,2X=140, 3 iso X=20 Une See Example 53 on Co X10 189, cy _ 15)'=215,2(X- 10)(¥=15) =60 ation (Ans.r=0.915, ¥=3X-27, ¥9=33] " eOmPUted through a conn ter: EX =125,5Y =100 yy2 ~ 650,372 « 46 Sey = son wa0s Later on it Was dise mee ‘OVered that . instead of (8, 12) and (6, got t™® Pais SE a Y ecomect » 8). Dete coefficient of correlation, mmine © miscopied as (6, 14) and (8, 6) Tegression equations (ii) correct Tans. (9 x= 9, $56¥42.776, Y= 0.8K, (i) r= 0.67] On each of 30 sets, ty FORK, (ii) r= 0. : EX=15,rY= © measurements are ‘made, The following summaries are given : cae SEAY=56, 5464 od 9¢ Calculate the product mo, ~ arithmetic means (XPPare equi ‘nt and the slope of regression line of ¥ on X ment correlation coeff [Hint: See Example 59) a aa : [Ans.r=0856,4x=1.10) ‘hen the method using actual values actual values, deviations taken from 1on process. In such a ‘ase, regression - In such ca Used to simpli llows: Regression Equation of Y on x 4 lations are expressed ag fo| LL bey or Y Ys 6yx(x—¥) — Here, X = Arithmetic mean of XS ¥ = Arithmetic mean of Y byx = Regression coefficient of Y on X Using deviations from actual means, the value of ‘byx can be calculated as: where, x =X -¥;y=y-7 Regression Equation of X on Y = ion coefficient of X on Y. where, bry = Regression coe! Using deviations from actual means, the value of bxy can be calculated as: where, y=J=2 = < ‘lear. The following examples make this method more cle% Business Sig, Example 9 — Obtain the two Tegression equations from the following data: x 2 4 8 8 ua 2 Y: 4 2 5 uJ s 6 Solution: Tes 7 * w “ a x : v -¥) x y 2 4 Py 4 a | oq 4 3 9 2 3 9 6 “1 1 5 0 0 8 1 1 10 +5 25 10 8 9 3 2 4 12 +5 25 6 +t 1 %=0 | x2=70 | x¥=30 | sy=0 By2=40 | Eay=1g Since, the actual means of X and Y are whole num! from X and Yto simplify calculations: By _18_ byx Sper 7g 0257 2 _18 bry “SP O45 Regression Equation of Y on X Y-¥=byx(x -X) ¥-5 =0.257(X-7) Y-5 =0.257X 1.799 . 0.257X + 3.201 Regression Equation of X on Y X-X= bry (¥-¥) X-7 =045(-5) ).45¥- 2.25 ASY-225+7 ASY + 4.75 bers, We should taken deviations‘Ate results Of the two series X and Y: 60, Exy =3900 "the respective means) By? 3g69 1363 Regression Equation ofX ony X-X= by (Y ~7) X-90= 1.363 (¥~79) X-90= 1363Y—95 4) 363Y~ 5.4) fY onx byx (X ~%) ¥~70 = 0.613 (v_ 90) Y~10 =0.613x~ 55.17 Y= 0.613x+ 14.93 Regression Equation o y- Example 11. The following table ives the aptitude test scores and Productivity indices of 10 workers at random: Antitude scor Productivity index 60 68 62 60 65, 62 70 80 72 85, 48 0 53 GO 3 a 65 = 82 81 Estimate: the test scot worker whose productivity index is 75 ity in re of a worker ) the productivity index of a worker whose test score is 92 (ii) the pr98 Solution: Calculation of Regression ee x | Productivity] (Y = 65) ¥ Nome a index y Score x a | x W 5 25 68 +3 9 62 3 9 60 ay 25 e o 0 62 3 a 0 +5 25 80 +15 225 oF 7 7 49 85 +20 400 ory + i 47 289 40 25 a in - 7 1 53 12 144 52 3 a oa 73 +8 64 62 - A “ 65 0 0 60 i Bs 3 81 +16 256 a 2v=650 | sy=0 | sy2=1752 |Zxy=10% 065 10 Regression Equation of X on ¥: X- ¥= bxy(¥- 7) = By 1044 «9 506 By? 1752 X-65 = 0.596 (Y ~ 65) X- 65 = 0.596 Y~38,74 or X= 26.26 + 0.596 Y Be Snding oa ea toe % of aperson whose productivity index (¥) is 75, put Xos= 26.26 +0.596(15) = 26.26444.7 = 79,96, Regression Equation of Y on X: ¥ - =byx(X ~X) dy BY = 104 gg Ex? 894 Y~ 65 = 1.168 (X-65) Y~65 = 1.168 X-75.92 or Y =-10,92 + 1.168 x For finding out the productivity index Put X = 92 in the above equation. Yop =-10.92 + 1.168(92) 10.92 +107.456 = 96.536 (Y) ofa worker whose test score ®)iso2 , we_—- segression Anal 99 PORTANT TYPICAL EXAMPLES aample 12+ solution: The following table show, Ss the Number, ‘of motor regist certain territory fora : ws the gistrations ina c as term of 5 years and the sale of Motor tyres by a firm in that territory for the same period: Year: 2 3 4 s Motor registration: 630 720 750 ey No. of tyres. I: tyres sold: 4,100 7,300 1,350 1,500 Find the regression e, ‘uation to estimate the sale of ‘Known. Estimate the sale of tyros wh f tyres when motor registration is hen registration is 850. LetX a number of motor Tegistrations and Y denotes the number of tyres sold To simplify the calculation, let X-X Spee | | = 1 +10 a 16 +28 0 0 0 +4 1 +8 +4 16 +40 By=0 | zy2=34 | Exy= 03 83 50 sex = 149; Bx? i, 27819 4978 Regression Equation of Y on X Y-¥=byx(X-¥) Y- 1300 = 1.4928 (¥-700) Y- 1300 = 1.4928x- 1044.96 Y = 1.4928X+ 255.04 The estimate of sale of tyres (¥) when registration X= 850 is given by Y= 1.4928 x 850 + 255.04 = 1268.88 + 255.04 = 1523.92 = 1524 since the number of tyres cannot be fractional. eea Busine 100 Si Example 13. Calculate the correlation coefficient from the following results: N=10,2¥ =350,2Y =310 £(X ~39)? =162,5(y ~31) =222,5 (¥ -35)(¥ -31)=92 Also find the regression line of ¥ on X, Solution: Thus, the given deviations (X35) and (¥-- 31) are from actual means (¥=35, ¥ =31), Thus, EX ~35)* =162 or Dx? =169 ZY ~31)? =222, or sy? =299 2 (4 ~35)(¥ -31)=92 or Yay =99 Coefficient of Correlation a a 0.485 Ex? fey? Vie Jo Regression Equation of Y on X Y-Y=byx (x -¥) Y-31= 0.568(¥- 35) Y-31=0.568x— 19.88 Y=0,568X~ 19. 88 +31 Y=0.568X+ 11.12 © Graphing Regression Lines Tt is quite easy to graph the regression lines once they have been computed, The procedure adopted is as follows: © Regression line of X on Y. The regression line of X on ¥ can be drawn with the help of regression equation of X on Y, i.e, X=a+by If we put the respective values of Y in the above regression equation, we win} find the estimated values of X. If we plot estimated values of X with th actual values of Y on 4 ‘graph, we can draw regression line of X on Y. the (@ Regression line of ¥ on X. The regression line of Y on X can be GW With the help of regression equation of Y on X, ie, Y=a+bxww Analysis Regression ine”! 1fwe put the respective values of X in the above equa of ¥. If we plot estimated values of Y with the actual values of X on the grap regression line of Y on X. ‘The following example illustrate the graphing of regression lines. imple 14. From the following data: esa (A) Obtain the two regression equations. (i) Draw up the two regression lines on the graph paper with the help of two regression equations. x: 1 2 3 cs 5 4 6 Calculation of Regression Equation ¥=5 y ¥ av 0 o eo =A 1 o +1 4 al By=0 | xy?=2 | ixy=+1 Regression Equation of Y on X Regression Equation of X on Y Y ~¥ =byx(X-X) X-K=by(Y-Y¥) 1 ¥ -5=-(X -2) oh ) 1 Y-5 36 ) y-s=lx-1 2 =ly+4 2 (ii) Regression Lines: In order to draw up the two regression lines on the graph, we shall have to plot the given values of X and the computed values of Y and the given values of Y and the computed values of X Computed Values of Y Regression equation of Y on X Regression equation of X on Y 1 1 Ly+4 X=iY-3 2 ‘Computed Values of X Yeee ee lived when V=1, Y= ()+4=45 when ¥=2, 7=4(2)44=50 when X'=3, P=! Q)+4255 uation forthe following data: Y (Dependent Variable) 18 12 10 ‘olution: x x 2 16 4 4 5 1 6 o 0 8 +2 4 it +5 25 2X=36 | tx=0 | oy2 259car Regression Analysis ne eg x i Regression Equation of Yonx eae Ys— 134x418. 04 EXERCISE 2.3 . For the following dat = 1 expenditure of ts ie 4P Fegression equation and ost imate sales for an advertisement leaks ea — Sales (¥ crore): 14 16 18 20 24 30 32 Adv. expenditure (& lakh): | 59 62 65, 70 76 80 78 (Hint: Let x denote sales) FAns.X=0.621Y-20.85, X,, = 25.125] Find the correlation coefficient andthe equation of X and Y: ofegression lines for the following values x: 14 7 2 5 8 6 10 aN 7 5 3 2 6 4 8 < (Ans. r= 0.884, X=0.75 +1 25Y, ¥ =0,6054 0.625x) 3. The following data relate to marketing expenditure and the corresponding sales: [Expenditure (X) (¥ lac): 19 12 18 20 B Sales (Y) ( crore) : 14 7 23 2 Ea Estimate the marketing expenditure to obtain a sales target of & 40 crore (Ans. X= 0.59Y + 3.02; X= 26.62] The following are the intermediate results of the two series X and Y X =65,Y =65, N =10,3x? =894,Zy? =1752,Exy =1044 : ‘ati tive means.) Where x and y are deviations from the respec . a two re; ae equations. Also estimate Y when X= 92 and X when Y=75, - i [Ans, Y= 1.168X 10.92, Yr = 96.536; X = 0.596Y+ 26.26, X75 ~ 70.96]Busines, 104 es Su, 5. An investigation in to the demand for Television sets in 7 towns has resulted in the. folly data: 7 [17 | 2 Population ('000) (x): 4 14 14 | 1 i a 1 3) No. of T.V. sets demanded (Y): | 15 27 27 30. 38 Calculate the regression equation of Y on X and estimate the demand for T.V, sets fora ty } with a population of 30 thousand. _ (Ans. ¥=—3 +24,» 4} 6. A departmental store gives in-service training to its salesmen which is followed by a tes. “considering whether it should terminate the service of any salesman who does Ot dowel the test. The following data give the test scores and sales made by nine salesmen trig, certain period: [Test scores: | 4 {19 | 24 | or [2 | 2 | 15s | 20 Es} [Sates (00 | 31 [36] ae Tap ot ram |inoaua | agai 2] Calculate the coefficient of correlation between the test scores and the sales. Does it indicate that the termination of services of low test scores is justified ? If the firm wants a minimyp sales volume of ™ 3,000, what is the minimum test score that will ensure continuation of service ? : .9471, justified, X= 14 429.14) 7. i gives the marks in Economics and Statistics of 10 students selected a 28 35 32 31 36 29 38 34 | 2] 4 {4 | 1 | 36 [a2 [at Pao 33 | 39 Find (i) The two regression equations, (@) The coetticient of correlation between marks in Economics and Statistics, (Gi) The most likely marks in statistics when marks in economics are 30, Ans. ().X=~ 0.23377+ 40,8806, y 0.6643 ¥ + 59.2576 (=~ 0394, (i) 39.3286, or 39 marks] 8. The profits (Y) of a company in the Xth year ofits life Were as follows: Years of life (X): 1 2 3 4 | e [Profits (¥) (in lakh of 2): 1250 1400 1650 1950 2300 Estimate the profit of a company in the 6th year. Ams. ¥=265%4 915, yg 2505 lk} 9. From the following data: (0 Obtain the two regression equations. (ii) Draw up two regression lines on the graph paper. [ x 65 66 67 [x 67 68 64ion Analy Linea 1s (@) Using Deviations taken from, Assume ses When actual means tum ou to be in Fa d Means ete then it becomes ditficul tg tay aviation her than the whole numbers like 24,69, 25,12 chaliculty, devitions fy Med meg tual means and squaring them up, To avoid sgson equations ae expe follows Mer than actu ments tru asl suich case, Regression Equation oy onx “ y-Fe byx OY % Regression eo : Rady — 2d Bay, byx = Lax N where, dk =X —4,, dy xy _ 4 ly Regression Equation of X on y X-X=py(y 7) Where, bxy= Regression Coefficient of X on y Using deviations from assumed means, the value of bry can be calculated as: buy = N *2dedy Bde. Day hy N .2dy? — (ay)? or Eddy — 2 Edy “ oH) Sdy? Cay” a N where, di=X—A,,dy=¥—A, The following examples will clarify this method. Example 16. Obtain the two Tegression equations for the following data: x 43 | 44 | 46 | 40 | 44 | 42 | 45 42 | 38 | 40 | 52 57 Y: | 29 | 31 19 | 18 | 19 | 27 | 27 | 29 | 41 | 30 | 26 | 10 Also find the value of X when Y=49 and Y when ¥=50. Hence or “otherwise find *y’~ Business Stig, q 106, ie tions Solution: Calculation of Regression Equa Aza] ax? Y A=27 dy’ ray x ix Fs dx 7 7 43 1 1 29 2 a 44 2 4 3t 4 as 8 46 4 16 19 38 es ~32 40 2 4 18 3 18 44 2 4 19 8 64 ~16 42 0 0 7 =A 0 0 0 45 3 9 ar 0 0 0 42=A 0 0 29 2 4 0 4 41 “4 196 56 30 3 9 -~6 26 - 1 =10 10 -17 289 ~255 PY=306 | rdy=-18 | say2= | Soxay= 728 347 7=a2t _306_ 55, NB Since the actual means of X and Y are in fractions, we should take deviations from assumed mean to simplify the calculations, byx = N xZdrdy ~Ede dy N .2dx? ~(Sdx)? = 12%(-347)~(29)(-18) _ ~41644529 ~3642 12x383—(29)? 4595-841 3755 =~ 0.969 =-0.97 bay = Eddy -Zadr Bay N Edy? (Say)? 212X347) -29)(~18) | 4164 +529 = 73642 12% 728 ~(-18)? 8736-324 “Fai =~ 0.432 =~ 0.43 Regression Equation of X on Y i Regression Equation ory =by(¥-7) aie X ~44,42=—0.43(¥ -25.5) X -44,42=~0.43Y +10.965 X =—0.43Y +55,385 onXyr 107 jon Analysis yxpar ROSES When Y = 49, When X = 50, X=-0.43Y + 55.395 Y =-0,97(50) + 68.5874 =~ 0.43 (49) + 55.385 — 48.5 + 68.5874 =~ 21.07 + 55.385 20.0874 » X49 =34.315 0.0874 Coefficient of Correlation r= Vbyx.bey (0.97) (0-43) =— 0.645 sample 17. Obtain the regression equation of Y on X from the following data: x: 78 89 97 69 59) 79 68 = ¥: 125 137 156 112 107 136 124 108 Calculation of Regression Equations x A=69 dx? Y is dy? dxdy ax 78 +9 81 125 169 +117 89 +20 400 137, 625 +500 97 +28 784 156 1936 +1232, 69=A oO oO M25A o oO co) 100 107, +5 25 +50 100 136, +24 576 +240, 1 124 42 144 ~i2 64 108 4 16 +32 Box? | EY= 1005 | zdy= 109 | say? Zdxdy = 1530 = 3491 = 2159 ee ~ = 125.625 N.2dxdy —Ede Sdy N Bde? — (Ede)? 8x2159~(48)(109) _ 17272-5232 _ 12040 © 8x1530-(48)? 122402304 ~ 9936 Regression Equation of Y on X Y-Y=byx(X-X) Y— 125.625 =1.212(X -75) Y— 125.625 = 1.212X-90.9 Y= 1.212X+ 34.725 = 1.212lf & = Business Sey & IMPORTANT TYPICAL EXAMPLES. Example 18. A panel ofjudges A and B graded seven independently and awarded the following a Debator: 1 2 3 4 u 6 7 ‘Marks by A: 40. 34 28 30. 44 38 [ay Marks by B: 32 39 26 30 38 34 An eight debator was awarded 36 marks by udge A while Judge B was not press the Judge B was also present, how many marks would you expect him to awa cighth debator assuming degree of relationship exists in judgement? Solution: Let marks awarded by Judge A be denoted by X and marks awarded by judge By denoted by Y. The marks expected to be awarded by Judge B can be determined by fitting regression equations of Y on X. Calculation of Regression Equations nt, If Id a A=30 dx? Y A=30 dy? dxdy dx. ay 40 +10 100 32 2 4 20 4 “4 16 39 9 81 36 28 -2 4 26 “4 16 eI 305A 0 0 302A 0 0 a 44 14 196 38 8 64 112 38 8 64 34 4 6 32 3t 1 1 28 5 i 5 Pox =35 | sax? =381 | EY = 207 Lay = 17 | say? = 495 Eaxay = 206 = NEdrdy ~Zdv .Zdy N Ede? —(Sdx)? _.7x206~(35)(17) "”7x381-(35)? | = A595 847 a so 2667-1225 1442 Regression Equation of Y on X Y-Y=byx(X-¥) ¥-32.43 =0.587 (¥-35) Y- 32.43 = 0,587 - 20,545 Y=0,587X+ 11.885587 (36) 4 Thus, if the Jug, debator, 9. Simple observations 9. - seample 1 8 Obtained t i and the length of the trouser 2° Measure of ‘the Wai Length of Trouser T1885 = 9 4 ge By, “1324 11.995 = : 33.017 of 33 approx. *S also present, he Would hav: ‘© awarded 33 marks to the eighth tudy the relation between the measure of the waist Are shown ag under : FS (In cm): Obtain the line of best fit i t re waist. Calculate the coef ctofan) £0 |825| 85 [87.5] 90 [ozs 105} 140] 95 | 98 [100 | 105 length of tro ient of, determination users on Measurement of the solution: Let X= measure of waist and Y= length of trousers, Here, *812.5,2Y= 1919 81.25 ang ince, X isnot st an wet a ueees He wil take the deviation of X fr The calculations are: x dx axe Y dy ay dxdy 70 =10 100 100 =1 1 10 72.5 - 75 56.25 102 +1 1 -75 75 -5 25 4100 “1 1 +5 775 -25 6.25 95 -6 36 +15 80=A 0 0 105 | 44 16 0 82.5 25 6.25 110 +9 81 + 225, 85 5 25 95 -6 36 -30 87.5 75 56.25 9 | -3 9 ~225 90 10 100 100 all 1 ~10 92.5 12.5 156.25 105 +4 16 +50 EX = 812.5] 2d = 12.5 lex? = 531.25] EY = 1010| xay =o Eay? = 198] Eddy = 32.5 Regression Coefficient of Y on X: ° N x2drdy —2dr dy DN de (it (10x32.5)-(12.5x0) 325 © (10x531.25)-(12.5)? 53125-15625 = —225_- 0.96 5156.25Business o Mes Stay L ig Line of regression of length of trousers on the measurement of the waist, Le, the lng of regression of Y on Xis _ Y-Y=byx(X-¥) Y-101 = 0.06 (X- 81.25) Y-101 = 0.06X- 4.875 Y =0.06X + 96.125 Coefficient of Determination: 2 pe N3dxdy -(Zdr).(Zdy) IN Bde? ~(Bde)? [NW Say? - Bay)? 2 =| 0«32.5)-(125%0) 4{10%531.25-(12.5)? /10%198-0 2 2 = 325 7 325 = ——| =| _ {5312.5-156.25 «1980 [5156.25 x/1980 = 625) __ 105625 _ ~ 5156.25x1980 "10200375 ~ Example 20. Fora bivariate data, you are given the following information: Z(X-58)=46 D(x 58)? = 3086 L(Y -58)=9 £(¥ ~58)* =493, 2(X ~58)(Y ~58) =1095 N=7 (Assumed means of X and series are both 58.) You ar required to determine (i) the two regression equati : ms and of correlation between X and Y series, (4) the coefficient Solution: Since the assumed means of X and Y series are both 58, we have. 086 ° 2dr = 46, Lae Xdy=9 Zdy? = 483 Eddy = 1095 N=7 yx — Baty “Bae By N Zdx? ~(Sdx)? _7*1095~(46)(9) "73086 —(46)? __ 1665-414 _ 7251 so = 037 21602-2116 19486bry N Lidvily N Xdy? _, Further, N Reeression Equation °fX ony; by Wr 7) ~¥ ¥~64.51=2 29¢y _ X=2.207 65, Regression Equation of Yonx: ¥-P=bye(x 7 59.29 =0370y_ 64,57) ¥°59.29 =0.37¥ 93 99) Ye Y=037X~23 99) 4 59,29 Y=037X+ 35.399 r= by Bx T= 2.20x0.37 =0,902 EXERCISE 2.4 Obtain the two regression equations for the following data: x 8 6 4 Li 5 3 | | 14 9 8 5 6 7 6 Also find the coefficient of correlation from the regression coefficients, (Ams.X=3.1 40.4 he following data: 2 Obiain the two Tegression equations from t = 223+ 0.685X; r= )523] Age of husband (x): | 18 19 | 20 | 21 | 22 23 24 25 26 Age of wite (Y): 7 18 | 18 | +8 | a9 19 | 20 | a Par 9f correlation from the regression coefficients cient o (Ans, Y= 0.47% 8.225, x= 1 997- 149,r=+0967] Also find the coeffi eeBusines 2 Sai, 3. The height of fathers and sons in inches are: aT 3 [ 7 ] Height of Fathers: | 65 | 66 | 68 | 69 | 71 a +2 ate |e) Height of Sons: 67 | 68 | 64 | 72 | 70 Estimate (i) the height of son if the height of the father is 64 inches, and (ii) the height of father if the height of son is 71 th s v 's coefficient of correlation between them, Also calculate the value of Spearman's ci {Ans () 6.18, (i) 69.2, (i) Re 04g 4. The age and blood pressure of 10 university teachers are: Age: s6_| 42 | 36 | 47 | 49 | 42 | 60 | 72 | 63 | 55 Blood Pressure: 1.147 | 125 | 118 | 128 | 145 | 140 | 155 | 160 | 149 150 () Find the correlation coefficient between age and blood pressure. Gi) Determine the least square regression equation of blood pressure on age. (id) Estimate the blood pressure of a teacher whose age is 45 years. (Hint: See Example 51] (Ams. r=0.89, Y= 1.114483.758, Yar= 133.708 ~ 134 5. The following table gives age (X) in years of cars and annual maintenance cost ) in hundred rupees: Xx 1 3 5 7 9 a] Y: 15 18 24 B 2 | Estimate the maintenance cost for a4 year old car after finding the regression equation, ; (Ans. Y = 0.95X + 15.05; Y,= 18.85) 6 Obtain the two regression equations from the following data: 4 5 6 8 " 12 10 8 rs 5 Verify that the coefficient of comelation is the geomenie mean of coefficients. . Of the two regression (Hint: See Example 50] [Ans. X = 15.024 ~9, ITIY; Y =~ 9, 929x + . 14.717; p= 7. Calculate from the following data: 17:r=~ 0,954) (Two regression equations (i) Coefficient of correlation (iif) Most likely value of X when Y = 10. a SS ss | se | oo fos | oe Tro Tae vr S| so | 48 | 60 | o | 4 | 65 [70 [a Ans. X= 9.403 + 0.8514Y, Y= 0.884 +0.992X, r-=o.919, x Lo] = 17.913)13 "Bressi, and Arithmer ion Eq ie AUations from Cour i NS of efficient , itandard Deviatior When the valueg ory, Xanay, : Nt Of Correlation, St SANRHIONS are expressed in the fo) ont 26 ang (D Regression Equation ofy 8 Marner onX "OFX and Y sereg are given, then regression Where, byy —). The following exanp) fs make the above Example 21. You are giy, said method more cl on the following in lear, formation: x Y Arithmetic mean: 5 [2 Standard deviations Seep Correlation coefficient T=07 (® Obtain two n ‘gression equations, (ii) Estimate Y when (iii) Estimate x when Given, X =5,7 =12,, = 2. 2.60, =3.6,r=0.7 ( Regression Equation of X on Y Solution: X-X==r S2(y_7) Sy Putting the values in the equation, we get X-$ =0.51 (Y-12) X-5=0.51Y-6.12 X=051Y-1.12 X=-1.124051¥ or14 ° S —(W-X) Sy Putting the values in the equation, we get y ~12=0,7x25(y_5) 26 97X -0.97x5 97X 4.85 .97X —4.85 412 .97X $7.15 Y=7.15+0,97" () Most tikely value of Y when X= 9 For this purpose, or We use regression of ¥ on X Y=7.15 +097" Putting X = 9 in the equation, we get Y=7. 15+ 0.9709) =7.15 +8.73 value of X when Y= 12 For tis purpose, we use regression of X ony X=- Li2+0.s1Y = 12 in the equation, we get X=-1124 0. 51(12) 1.12+612=5 = 15.88 (ii) Most likely Putting Y Example 22. You are given below the following information about advertisement and sales: Adv. Expenditure Sales (crore) (¥ crore) Mean 20 SD. 120 5 |: Correlation costicient,ra= + 0.8 (@ Calculate the two Tegression equations, (4) What should be the advertisement budget if the company ants fo atain sales target of & 150 crore? (Gi) Find the most likely sales when advertisement expenditure is & 25 crore, olution: Let Y= Ady, Expenditure and Y= Sales Yy Thus, we have ¥ =20, =120,6, =5,0, = 25,1, =0.8wression Analy 1 is a aasnaeOnT CEE : ~ X=0. (6) Regression Eq ation of ¥ on x Yer Soy 5) Sy ¥-120=08 2am ¥ 120 =4 (29) Y- 120 =4y_ 9 Y=40+4y (ii) When sales target (Y) is % 150 crore, X=0.8+ 0.16 Put Y= 150, X=08 +0, 16(150) = 0.8 +24 = 24.8 crore, (iii) When advertisement expenditure (X) is % 25 crore, t Y=40+4y Put Y=25, y=40+ 4(25) =40+100= 140 crore, rample 23. Find the regression equations when you knows then the advertisement expenditure (X) is he sales (¥) is ¥ =682, 7 =99, 2 -044,r=076 Sshtion: Given, ¥ =682, 7 44,7 =0.76 (@ Regression Equation of Y on X: o(x-¥) 6, Putting tion, we get ‘utting the values in the equat Y -9.9=0.76x0.44(X -68.2)Business Sai 16. Y -9.9=0.3344(X -68.2) Y -9.9=0.3344X -22.81 ¥ =0.3344X —22.814+9.9 or Y =0.3344X -12.91 (i) Regression Equation of X on Y: Sy -7) oy X- 68.2 =0.76x2.27(¥ -9.9) X-68.2 =1.725(¥ -9.9) X- 68.2 =1.725Y -17.08 X=1.725Y -17.08+68.2 T2SY +51.12 Example 24. Find the expected price in Mumbai when price in Calcutta is € 70 using the following data: Average Price in Calcutta 765 Average Price in Mumbai 67 S.D. of Price in Calcutta 25 a5) S.D. of Price in Mumbai Correlation coefficient between price of Mumbai and Calcutta : 08 Let X = Price in Calcutta, Y = Price in Mumbai Given: ¥ =65,¥ =67,5, =2.5,0, =3.5,r=08 Expected Price in Mumbai (Y) when price in Caleutta (X) = 70 can be found from regression equation of Y on X. Regression Equation of Y on X Solution: Putting the values, we get 3.5 —67 =0.8x——(X -65) 7 25° Y -67=1.12(X -65)7 an Y-61=11ox-ng Ys112x 72.3467 12K 5,8 -12(70)~5.8=78.4-58 26 Thus, the expected price in Mumbai is ¢ 72.6 corresponding to % 70 at Calcutta, soap 25. The’ coofiient of correlation between the ages of husbands and wives in a community 0 48, the average ofhusband age was 5 years and that of wives age 22 years. Their standard deviations were 4 and S years respectively. Find with the help of regression equations: (0 the expected age of husband when wife's age is 20 years and (Gi) the expected age of wife when husband's age is 33 yeas, suation: Let age of wife be denoted by Y and age of husband by X. We are given: ¥=25,¥ =22,6, =4,0, =5,r=08 (0 For estimating age of husband when wife's ‘age is 20 years, we use regression of X on ¥ as follows: when X= 70, X-X=r “y-F) ° X-25=08% 4-29) X -25=0.64(Y -22) X -25=0.64Y -14.08 X =0.64Y +10,92 0, X= 0.64 (20) +1092 = 12.8 + 10.92 = 23.72 Thus the expected age of husband when wife's age is 20 years shall be 23.72 years (i) For estimating age of wife when husband’ age is 33 years, we use regression equation of Y on X as follows: when Y Ur) s, y- ¥~22=0.8%5 (4-25) 4 Y-22=1(X ) es when X= 33, Y=33-3=30 Thus, the expected age of wife when husband's age is 33 is 30 years u eeif Business Seu us ; IMPORTANT TYPICAL EXAMPLES ee marks ji Example 26, The following data based on 450 students are given for marks in Statistics, Feonomies at a cer ‘tain E amination: Mean Marks in Statsties : 40 Mean Ma : 48 SD. of Man sin Stati ; 12 The variance of marks in Economi 256 Sum of the product 's of deviation of marks from their respective means is 4207s (© Obtain the equations of two lines of regression (ii) Estimate the 'verage marks in Economics of candidates who obtained 50 ‘atk in Statistics, Solution: (i) Let X denote marks in Stati given: = 40, Y=48 9,=12, 6s and Y denote marks in Economics, We ay Ey =42075 Before we obtain the regression equ: ‘ations, we compute the Coefficient of Correlation (r) by using the formula. 450% 12x16 = +0.49 approx, Regression Equation of X on y X-Xar Sey py oLinear Regression Analysis Example 27. Solution: Example 28, Solution: ¥~48=0.653(x 40) ¥~48=0.653x ~26.12 653X ~26.12+48 ¥ =0.653x 421.88 (i) To estimate the marks in Economies when 50 marks in Statistics is given, we use regression of Y on X. Y=0.653X +21.88 When X= 50, Y= 0.653 (50) + 21.88 = 32.65 +21.88 = 54.53 or 55 marks Thus, the expected marks in Economics ig $5 1X =25, ¥ =120,bxy =2 Estimate the value of X when Y Given, ¥ =25, ¥ =120,bxy =2 130. For estimating X when Y= 130, we use regression equation of Xon Yas follows: X-X=by(y -¥) or X=X+by(¥ -7) X =25+2(130-120) X =25+2(10)=45 Thus, the value of is 45 when Y= 130 Ifo? =9, 0; =1600,r,, =0.5, obtain bxy. Given, of =9(or 6, =3), 0? =1600(or 6 =40),r, =0.5, We know ,I~ jon coefficient of Yon A, Le V 100 sag y =) 1 find bru 3 Example 20. Solution , Ura >} ) wV-¥) 22100 vets ¥=2u +100 => — ¥=3v4100 ¥=2u +100 7 =3v +100 (YT) =Au-a) . (P-1)=30-¥) now yy = 223 HU = O=V) 2 Sou 4-Su-u)? bry = OE " 15 ‘Mean marks in English ‘Mean marks in Maths | S.D. of marks in English Ke S.D. of marks in Maths Coefficient of correlation between English and Maths = (i) Obtain the two regression equations. (i) Calculate the expected average marks in Maths of candidates who received 50 marks in [Ans. X= 0.268Y+ 26.73, Y= 0.657X+21.64, 54.5] English 2. There are two series of index numbers, P for price index and S for stock commodities. The mean and standard deviation of P are 100 and 8 respectively and S are 103 and 4. The correlation coefficient between the two series is 0.4. With this data work out a linear equation to read off values of P for various values of S. Can the same equation be used to read off the values of S for different values of P? If not, give the appropriate equation, (Ans, P= 085+ 176; No; $= 02 + 83]y- ay Regression Analysis, pst 121 The following results were worked out from matks in Statistics and Mathematics in a certain examination: Marks in Statistics Marks in Maths « ) AM. 395 475 Se 108 178 Coefficient of correlation = 0.42 (i) Find the two regression equations, (ii) Estimate the value of Y when X= 50 and X, when ¥= 30, (Ans. () ¥=0.6922x + 20.1581, X= 0.2548Y + 27.397; (ii) 54.76, 35.04] 4, Youare given the following data about sales and advertise sment expenditure of a firm: Sales Adv. Expenditure (8 crore) (€ crore) Arithmetic Mean: 50 10 ‘Standard Deviation: 10 2 ‘Coefficient of correlation = +0.9 (i Calculate the two regression equations. (ii) Estimate the likely sales for a proposed advertisement expenditure of & 13.5 crore. (iii) What should be the advertisement budget if the company wants to achieve assales target of & 70 crore? [Ans. () X=4.5Y + 5, ¥=0.18X+ 1, (i) 65.75 crores, iif) 13.6 crores) Given the following data, what would be the possible yield when rainfall is 29"? Rainfall Yield per acre Mean: 25" 40 Variance: o 36 Coefficient of correlation between rainfall and production = 0.8 [Ans. 46.4] 6. Fora given set of bivariate data, the following results were obtained: ¥ =53.2,¥ =27.9, regression coefficient of Yon X=-1.5 Regression coefficient of X on Y= 0.2, Find the most probable value of ¥ when X= 60. ‘Also find the coefficient of correlation, [Ans, ¥op=17.7,r=-0.548] 1. WX =45, 0, =25, ¥=60,6,=22, r Estimate (i) Value of Y when X= 35 (ii) Value of X when Y= 20, 15 [Ans. (53.4, (i) 10.92]