Statistical Computing and R Programming V Semester BCA, Correlation and Regression Correlation Meaning: © Correlation is a statistical technique to ascertain the association or relationship between two or more variables. * Correlation analysis is a statistical technique to study the degree and direction of relationship between two or more variables. * A correlation coefficient is a statistical measure of the degree to which changes to the value of one variable predict change to the value of another. When the fluctuation of one variable reliably predicts a similar fluctuation in another variable, there’s often a tendency to think that means that the change in one causes the change in the other. Uses of correlations: 1. Correlation analysis helps inn deriving precisely the degree and the direction of such relationship. 2. The effect of correlation is to reduce the range of uncertainty of our prediction. The prediction based _ on correlation analysis will be more reliable and near to reality. 3. The measure of coefficient of correlation is a relative measure of change. ‘Types of Correlation: Correlation is described or classified in several different ways. Three of the most important are I. Positive and Negative IL. Simple, Partial and Multiple JIL Linear and non-linear 43Statistical Computing and R Programming V Semester BCA, I. Positive, Negative and Zero Correlatios Whether correlation is positive (direct) or negative (in-versa) would depend upon the direction of change of the variable, Positive Correlation: If both the variables vary in the same direction, correlation is said to be positive. It means if one variable is increasing, the other on an average is also increasing or if one variable is decreasing, the other on an average is also deceasing, then the correlation is said to be positive correlation. For example, the correlation between heights and weights of a group of persons is a positive correlation, Height (m):X | 156 | 160 | 163 | 166 | 168 | 171 | 174 | 176 Weight (kg):¥ [60 | 62 | 64 | 65 | 67 | 69 [ 71 | 72 Negative Correlation: If both the variables vary in opposite direction, the correlation is said to be negative. If means if one variable increases, but the other variable decreases or if one variable decreases, but the other variable increases, then the correlation is said to be negative correlation. For example, the correlation between the price of a product and its demand is a negative correlation, Price of Product (Rs PerUnit):X| 6 | 5 | 4 | 3 | 2 | 4 Demand (In Units) :¥ 75_| i120 [175 | 250 [215 [ 400 Zero Correlation: Actually it is not a type of correlation but still it is called as zero or no correlation, When we don’t find any relationship between the variables then, it is said to be zero correlation, It means a change in value of one variable doesn’t influence or change the value of other variable. For example, the correlation between weight of person and intelligence is a zero or no correlation.Statistical Computing and R Programming V Semester BCA, II. Simple, Partial and Multiple Correlations: The distinction between simple, partial and multiple correlation is based upon the number of variables studied. Simple Correlation: When only two variables are studied, it is a case of simple correlation. For example, when one studies relationship between the marks secured by student and the attendance of student in class, it is a problem of simple correlation. Partial Correlation: In case of partial correlation one studies three or more variables but considers only two variables to be influencing each other and the effect of other influencing variables being held constant, For example, in above example of relationship between student marks and attendance, the other variable influencing such as effective teaching of teacher, use of teaching aid like computer, smart board etc are assumed to be constant. Multiple Correlations: When three or more variables are studied, it is a case of multiple correlation. For example, in above example if study covers the relationship between student marks, attendance of students, effectiveness of teacher, use of teaching aids ete, it is a case of multiple correlation. III. Linear and Non-linear Correlation: Depending upon the cons! incy of the ratio of change between the variables, the correlation may be Linear or Non-linear Correlation. ear Correlation: If the amount of change in one variable bears a constant ratio to the amount of change in the other variable, then correlation is said to be linear. If such variables are plotted on a graph paper all the plotted points would fall on a straight line. For example: If it is assumed that, to produce one unit of finished product we need 10 units of raw materials, then subsequently to produce 2 units of finished product we need double of the one unit. 45Statistical Computing and R Programming V Semester BCA, Raw material : X 10 20 30 40 50 60 Finished Product : ¥ 2 4 6 a 10 12 Non-linear Correlation: If the amount of change in one variable does not bear a constant ratio to the amount of change to the other variable, then correlation is said to be non-linear. If such variables are plotted on a graph, the points would fall on a curve and not on a straight line. For example, if we double the amount of advertisement expenditure, then sales volume would not necessarily be doubled. ADVERTISEMENT 10 20 «| 30 40 | 50 EXPENSES: X SALES VOLUME: Y 2 5 8 4 12 Mlustration 01: State in each case whether there is (a) Positive Correlation (b) Negative Correlation (¢) No Correlation SINo Particulars Solution 1_| Price of commodity and its demand Negative 2__| Yield of crop and amount of rainfall Positive 3 __| No of fruits eaten and hungry of a person Negative 4 | No of units produced and fixed cost per unit Negative 5 _| No of girls in the class and marks of bos No Correlation 6 __| Ages of Husbands and wife Positive 7_| Temperature and sale of woollen garments Negative ‘8 _| Number of cows and mille produced Positive ‘9 | Weight of person and intelligence No Correlation 10 _| Advertisement expenditure and sales volume Positive Methods of measurement of correlation: Quantification of the relationship between variables is very essential to take the benefit of study of correlation. For this, we find there are various methods of measurement of correlation, which can be represented as given below: 46Statistical Computing and R Programming V Semester BCA, Methods of Measurement of Correlation Graphic Method Algebric Method 1. Karl Pearson's Coefficient of Correlation 2. Spearman's Rank Coefficient of Correlation 3. Concurrent Deviation Method 4, Method of Least Square 1 Scatter Diagram 2. Graph Method Among these methods we will discuss only the following methods: 1, Scatter Diagram Scatter Diagram: This is graphic method of measurement of correlation. It is a diagrammatic representation of bivariate data to ascertain the relationship between two variables. Under this method the given data are plotted on a graph paper in the form of dot. i.e. for each pair of X and Y values we put dots and thus obtain many points as the number of observations. Usually an independent variable is shown on the X-axis whereas the dependent variable is shown on the Y-axis. Once the values are plotted on the graph it reveals the type of the correlation between variable X and Y. A scatter diagram reveals whether the movements in one series are associated with those in the other series + Perfect Positive Correlation: In this case, the points will form on a straight line rising from the lower left hand corner to the upper right hand comer. + Perfect Negative Correlation: In this case, the points will form on a straight line declining from the upper left hand comer to the lower right hand corner. 47Statistical Computing and R Programming V Semester BCA, + High Degree of Positive Correlation: In this case, the plotted points fall in a narrow band, wherein points show a rising tendency from the lower left hand comer to the upper right hand comer. + High Degree of Negative Correlation: In this case, the plotted points fall in a narrow band, wherein points show a declining tendency from upper left hand comer to the lower right hand corner. + Low Degree of Positive Correlation: If the points are widely scattered over the diagrams, wherein points are rising from the left hand comer to the upper right hand corner. + Low Degree of Negative Correlation: If the points are widely scattered over the diagrams, wherein points are declining from the upper left hand corner to the lower right hand corn + Zero (No) Correlation: When plotted points are scattered over the graph haphazardly, then it indicate that there is no correlation or zero correlation between two variables. 48Statistical Computing and R Programming Perfect Positive Correlation V Semester BCA, Porfect Nogative Corrolation Diagram - 1 High Positive Correlation Diagram - 11 High Negative Correlation Diagram - 111 49 Diagram -1VStatistical Computing and R Programming V Semester BCA Loy Postive Correlation Low Negative Correlation Diagram-V No Correlation lustration 02: Given the following pairs of values: Capital Employed (Rs.InCrore) [1 [2[3[4]5 [7 [8] [i[i2 Profit (Rs. In Lakhs) 3{s[4{[71[91e [10[ai [a2 [14 (a)Draw a scatter diagram (b)Do you think that there is any correlation between profits and capital employed? Is it positive or negative? Is it high or low? Solution: From the observation of scatter diagram we can say that the variables are positively correlated. In the diagram the points trend toward upward rising from the lower left hand comer to the upper right hand comer, hence it is positive correlation. Plotted points are in narrow band which indicates that it is a case of high degree of positive correlation.Statistical Computing and R Programming an _ mt © ae oo 2]. M< Capital Employed (Rs. fn Crore) Correlation coefficient DIRECT METHOD Sev fo > Interpretation of coefficient of correlation A Negative value of 7 indicat Wawa y=0 then the variables are non-correlated. eon DIRECT METHOD ‘A positive value of y indicates positive correlation. s negative correlation. If y =+lthen the correlation is perfect positive If y =-Ithen the correlation is perfect negative. If y >=0.5 then correlation will be high degree of positive. If'y >=-0.5 then correlation will be high degree of negative If y <0.5 then correlation will be Low degree of positive, If y<-0.5 then correlation will be low degree of negative V Semester BCA, From following information find the correlation coefficient between advertisement expenses and sales volume using Karl Pearson’s coefficient of correlation method. Firm 1[2]3]4|5 9 [10 Advertisement Exp. (Rs. In Lakhs) | 11 | 13 | 14 | 16 | 16 13 | 13 Sales Volume (Rs. In Lakhs) 50 [50 [55 | 60 | 65 60 | 50 s1Statistical Computing and R Programming V Semester BCA Caleulation of Karl Pearson's coefficient of correlation Yblewew sea ooee 0 Bei + asa "ae [xamples on Karl Pearson’s coefficient of correlation : PRODUCT MOMENT COEFFICIENT OF CORRELATION PEARSON PRODUCT- (Sx MOMENT CORRELATION Sy »_ | — QOEFFICIENT (12.1) (3x7 1" Product moment coefficient of correlation is, nbxy ~ (2x)(2y) [ndx? - (2x)*][ndy? - (2y)?] Yy =Statistical Computing and R Programming V Semester BCA Calculate the Karl Pearson’s product moment of coefficient of correlation. Student 1 2 3 4 3 6 Statisties(x) [7 4 6 9 3 8 Mathematies(y) [8 3 4 8 3 6 olution: ~ * > y [=| | 7 | 8 | 49 | oF | 56 4 5 16 25. 20 6 4 36 16 24 9 [8 | a1 | o | 72 $s ePerts 8 | 6 | 64 | 36 | 48 37 | 34 | 255 | 214 | 229 Product moment coefficient of correlation is, ndxy — (2x) @y) s Bom = Gy ai Bnd ~ [@eni2s; Ye = 0.8081 ‘There exists a positive correlation of higher degree between xand y. 53Statistical Computing and R Programming V Semester BCA, equa Fats Computation ofr forthe Interest Index Economics Example —_Day x , ae fr y 1 1 al ARH 16803 2 ne mn oe 135 3 400 2% sins sion 4 1 as sas 17823 3 70 my S76 m4) ‘ 18 a 99 170.49 7 68 2B 9n9 17164 3 1 a sins 1a 9 138 26 51076 ATS ty) aor a 3525 18%4 u ft 3 5489 13099 2 san ML esata san 1st800 Sxe9053 Syem9a7 Soy=21,0507 Correlation mat A correlation matrix is a table showing correlation coefficients between variables. Each cell in the table shows the correlation between two variables. A correlation matrix is used to summarize data, as an input into a more advanced analysis, and as a diagnostic for advanced analyses ‘An example of a correlation matrix Typically, a correlation matrix is “square”, with the same variables shown in the rows and columns. I've shown an example below. This shows correlations between the stated importance of various things to people. The line of 1.00s going from the top left to the bottom right is the main diagonal, which shows that each variable always perfectly 54Statistical Computing and R Programming V Semester BCA, correlates with itself. This matrix is symmetrical, with the same correlation is shown above the main diagonal being a mirror image of those below the main diagonal. Y Xt X2 The above table is a correlation matrix between different Bonds issued by the Government with different residual maturity stated in the form of years in both horizontal and vertical buckets. It enables us to interpret that a bond with 0.25 years to maturity and a bond with 0.5 years to maturity has a correlation coefficient of 0.97 in their price movements and similarly for other maturity bonds. Maturity Buckets {In Years) oas[oso| 1 | 2[ 3 [5 | 10] 45 | 20] 30 oso] o97| 4 1 [ost|oa7| + 2 | oai [ost|os7| a 2 [072[086/094|000) a 5 | 057 |076) 0.89 096/028] 4 10 | 040 0.57|076|0.89| 0.93 |os7| 1 15 [040 [0.42| 0.55 |0.82| 0.89 [084/099] 1 20 [0.40 0.40/057 0.76] 0.24[o.01[0.97/099| 4 30 | 0.40 | 0.40] 0.42 [0.66] 0.76 [0.26] 0.94 [0.97 |o99| 1 55Statistical Computing and R Programming V Semester BCA, Applications of a correlation matrix There are three broad reasons for computing a correlation matrix: . To summarize a large amount of data where the goal is to see patterns. In our example above, the observable pattern is that all the variables highly correlate with each other. To input into other analyses. For example, people commonly use correlation matrixes as inputs for exploratory factor analysis, confirmatory factor analysis, structural equation models, and linear regression when excluding missing values pairwise. As a diagnostic when checking other analyses. For example, with linear regression, a high amount of correlations suggests that the linear regression estimates will be unreliable. REGRESSION Meaning: Regression analysis is a statistical tool to study the nature and extent of functional relationship between two or more variables and to estimate (or predict) the unknown values of dependent variable from the known values of independent variable. The variable that forms the basis for predicting another variable is known as the Independent Variable and the variable that is predicted is known as dependent variable. For example, if we know that two variables price (X) and demand (Y) are closely related we can find out the most probable value of X for a given value of Y or the most probable value of Y for a given value of X. Similarly, if we know that the amount of tax and the rise in the price of a commodity are closely related, we can find out the expected price for a certain amount of tax levy. Uses of Regression Analysis: 1. It provides estimates of values of the dependent variables from values of independent variables. 2. Itis used to obtain a measure of the error involved in using the regression line as a basis for estimation. 3. With the help of regression analysis, we can obtain a measure of degree of association or correlation that exists between the two variables. 4, It is highly valuable tool in economies and business research, since most of the problems of the economic analysis are based on cause and effect relationship. 56Statistical Computing and R Programming V Semester BCA, ear Regression Regression lines and regression equations are used synonymously. Regression equations are algebraic expression of the regression lines. Let us consider two variables: X & Y. If y depends on x, then the result comes in the form of simple regression. If we take the case of two variable X and Y, we shall have two regression lines as the regression line of X on Y and regression line of Y on X. The regression line of Y on X gives the most probable value of Y for given value of X and the regression line of X on Y given the most probable value of X for given value of Y. Thus, we have two regression lines. However, when there is either perfect positive or perfect negative correlation between the two variables, the two regression line will coincide, i.e. we will have one line. If the variables are independent, r is zero and the lines of regression are at right angles i.e. parallel to X axis and Y axis. Therefore, with the help of simple linear regression model we have the following two regression lines 1. Regression line of Y on X: This line gives the probable value of Y (Dependent variable) for any given value of X (Independent variable). Regression line of Y on X 2 Y-Y= bye (X-X) OR : Y=a+bX 2 Regression line of X on Y: This line gives the probable value of X (Dependent variable) for any given value of Y (Independent variable). Regression line of Xon Y HRs by OR X=a+bY In the above two regression lines or regression equations, there are two regression parameters, which are “a” and “b”. Here “a” is unknown constant and “b” which is also denoted as “by,” or “by”, is also another unknown constant popularly called as regression coefficient. Hence, these “a” and “b” are two unknown constants (fixed numerical values) which determine the position of the line completely. If the value of either or both of them is changed, another line is determined. The parameter “a” determines the level of the fitted line (i.e. 87Statistical Computing and R Programming V Semester BCA, the distance of the line directly above or below the origin). The parameter “b” determines the slope of the line (i.e. the change in Y for unit change in X). If the values of constants “a” and “b” are obtained, the line is completely determined. But the question is how to obtain these values. The answer is provided by the method of least squares. With the little algebra and differential calculus, it can be shown that the following two normal equations, if solved simultaneously, will yield the values of the parameters “a” and “b”. Two normal equations: Xon¥ Yonx DX = Na+byy DY = Na+bpx SExy = abY+byye rxy aX + DEX? This above method is popularly known as direct method, which becomes quite cumbersome when the values of X and Y are large. This work can be simplified if instead of dealing with actual values of X and Y, we take the deviations of X and Y series from their respective means. In that case: Regression equation Y on X: a+bX will change to (¥-¥) = by (X-X) Regression equation X on ¥: =a+bY will change to (X-X)=be (Y-) In this new form of regression equation, we need to compute only one parameter i.e. “b”. This “b” which is also denoted either “by.” or “bs,” which is called as regression coefficient. Iustration 0 Find the two regression equation of X on Y and Y on X from the following data: m to 12 16 11 15 14 20 22 15 18 23 14 20 17 25 28 58Statistical Computing and R Programming V Semester BCA, Calculation of Regression Equation = ¥ ma ye xy, 10 15 100 225 150 12 18 144 324 216 16 23 256 529 368 4 14 a2a 196 154 45 20 225 400 300 14 7 196 289 238 20 25 400 625 500 22 28 484 784 616 120 160 1.926 3,372 2,542 EX xY 5x" ry ExY. Here N = Number of elements in either series X or series Y = 8 Now we will proceed to compute regression equations using normal equations. Regression equation ofXonY: X=a+bY ‘The two normal equations are: DX = Na+byy DXY = 9 ayy+byyve ‘Substituting the values in above normal equations, we get w20 = 8a + 160b 2542 = 160a + ~—-3372b Let us solve these equations (i) and (ii) by simultaneous equation method Multiply equation (i) by 20 we get_ 2400 = 160a + 3200b Now rewriting these equations: 2400 = i + 3200b 2542 = 0a + = -3372b “142 = -172b ‘Therefore now we have -142 = -172b, this can rewritten as 172b = 142 Now, b =1# = 0.8256 (rounded off) Substituting the value of b in equation (i), we get 120 = 8a + (160 * 0.8256) 120 = 8a + 132 (rounded off) Ba = 120 - 132 Ba = “12 a = -12/8 a s “15 ‘Thus we got the values of a = -1.5 and b = 0.8256 Hence the required regression equation of X on ¥: Xsa+bY => .5 + 0.8256Y 59Statistical Computing and R Programming V Semester BCA Regression equation of ¥ on X: Ysa+bX ‘The two normal equations are: DY = Na+byx DXY = alX+byxX? ‘Substituting the values in above normal equations, we get 160 8a + — 120b o~- (iii) 2542 120a + = —-1926b Gv) Let us solve these equations (iii) and (iv) by simultaneous equation method Multiply equation (iii) by 15 we get 2400 = 120a + 1800b Now rewriting these equations 2400 12a + 1800b 2542 = 0a + 1926 142 -126b ‘Therefore now we have -142 = -126b, this can rewritten as 126b = 142 Now, b A = 1.127 (rounded off) Substituting the value of b in equation (iii), we get 160 = 8a + (120*1127) 160 = Ba + 135.24 Ba = 160 - 135.24 Ba 24.76 a 7 24.76/8 a 3.095 Thus we got the values of a = 3.095 and b = 1.127 Hence the required regression equation of Y on X: Ysa+bX => Y=3.095+41.127X Mustrs 02: Compute the regression equation of y on x from the following data. x 2 4 5 6 8 i Y 5 60Statistical Computing and R Programming V Semester BCA Solution: x y x xy 2 18 4 36 4 12 16 48 5 10 25 50 6 8 36 48 8 7 o4 56 a 5 121 55 36 60 «| 266 | 293 sg ee iggy vee” pany. @ ya = = 10 _ndxy- ExEy_ 6 x293-36 x60_ —402 Be = x)?” 6 * 266 — (36)? 300, es Regression equation of y on x is, Y- P= bye - %) i.e. (y— 10) = -1.3333(x- 6) i.e., y= -1.3333x + 7.9998 + 10 iLe., y = -1.3333x +17.9998 ind the regression equation of x on y and predict the value of x when y is 9. x 3 6 5 4 4 6 7 5 ¥ 3 2 3 5 3 6 6 4 Solution: x | 3 6 s[4]4¢[6][7]5 [Ex= 40] y | 3 al s[ sas] s.[ 6.6) 4 [ty 32 2 (eo o | 25| 9 | 36 | 36| 16 |=y7— 144] xy | 9 [12] 15 | 20 | 12 | 36 | 42 | 20 [Say — 166 61Statistical Computing and R Programming x_ 40 eo oe 7* ga ey 2 z a ae _ nny ExEy _ 8x 166-40 x 32 ‘v noy?— @y! . 8x 144— G2)? Regression equation of x ony is, &=- 9 = by- 7) Le, (x~ 5) = 0.375 (y-4) x=0375y-1545 ie, x=0.375y + 3.5 Wheny=9, x =0.375*x9+3.5=6875=7 Ilustrat 04: Find the two regression lines from the following data. V Semester BCA, 61 62 a x 55 57 58 59 59 60, Y 74 [77 [78 [75 [78 “(82 82 79. 81 Solution: x x xt # xy ao 74 | 3025 | 5476 | 4070 37 77 | 3249 | 5929 | 4389 58 78 | 3364 | 6084 | 4524 59 75 | 3481 | 5625 | 4425 59 78 | 3481 | 6084 | 4602 60 82 | 3600 | 6724 | 4920 61 82 | 3721 | 6724 | 5002 62 79 | 3844 | 6241 | 4898 64. 81_| 4006 | 6561 | 5184 335 | 706 [31861 | 55448 | 42014 62Statistical Computing and R Programming V Semester BCA, 9% 42014 ~535 x 706 "9x 55448 — (706)? Hoe nDxy- DxEy _ 9x 42014—535 x 706 eee Ca 9x 31861 — (535)? Regression equation of x on y is, @-D= by -— Le,, (x ~ 59.4444) = 0.698 (y - 78.4444) ie., x = 0.698y ~ 54.7542 + 59.4444 0.698 = 0.7939 ie,, x = 0,698y + 4.6902 Regression equation of y on xis, G- Y= by &- XY) ie, (y ~ 79.4444) = 0.7939(x - 59.4444) ie, y = 0.7939x - 47.1929 + 78.4444 -ie.,y = 0.7939x + 31.2515 63