5.7___Curve Fitting The set of data obtained from measurement or observations can be used for further study e the data collected from a group of people by measuring the heights of som and father can be used for finding the relation between the heights of son and father. The data collected can give us different types of approximate equations or relations. But the problem is to find the equation of the curve of best fit which can be used for prediction. The process of finding the equation of best fit is called curve fitting. The curve fitting can be used to study the relationship between the variables like height and weight or age and blood pressure of peoples etc. ‘There are different methods available to fit the curve like graphical method, method of moments and method of least square etc. The method which is more widely used and best to fit the curve is least square. It is very easy to implement on computer also. We will learn how to fit a straight line and a second-degree curve with the help of method of least parabolic squares.Ex. 6.7.12 Fit a straight line for the following data sla[sle|[7|s5 elelsl[r7lela And estimate y when x = 2 son. Let y = a + bx be a first-degree curve ie. a straight line. —__———_—. ‘Lo fit this straight line, the normal equations are = anal. = » = l= 3 8 24 2 4 6 24 36 5 S 25 26 é z 42 a6 7 6 42 49 S 4 se | 6a === ss | =y— 36 | Sey — 189 | Sx — 199 qf ° a zSubstitute the values of the ‘sums obtained from table in normal equations, The normal equations become 36 = 6a+33b 189 = 33a+199b Solving above equations simultaneously for a and b we get, Wl a = 8.8286, b =— 0.5143 Put the values of a and b in y = a + bx Therefore, we get the required straight line as Y = 8.8286 — 0.5143x. ‘When x 2, from y = 8.8286 — 0.5143x y¥_ = 8.8286 — (0.5143) (2) = 7.8Ex. 6.7.2 : In @ study between the amount of rainfall GO and the quantity of air pollution () removed the following dats Were collected Fit a straight line for the following data. [= a3 [ss [40 [46 [54 a2 |]26 [14 vilase[a11 140.0] 10.01 10.4] 10.8 | 12.2 [19.4 of alr pollution y when x= 10.6. and estimate the quantity Son. : Let y = a + bx be a first-degree curve Le. straight line. To fit this straight Sy = Na+bex Say - abet b Ex? We construct a table to find the values of constants line, the normal equations are a and b. = ¥ ~ = ae 3.3 416 sR.28 10.89 [3.6 43.4 35.85 12.25 4.9 10.6 61.04 24.01 4.6 10.8 49.65, 21.16 6.1. 10.4, 53.04 26.03 4.2 10.8 17.64 25 12.2 34.16 754 4 13.4 44.44 2.21 2x = 29.5| sy = 90.6 = 825.72 121.01 Here N=5Substitute the values of the sums obtained from table in normal equations, The normal equations become 90.6 = 8a+29.5b 325.72 = 29.5a + 121.01b Solving above equations simultaneously for a and b we get, a = 13.8482, b = — 0.6842 Put the values of a and b in y = a + bx. Therefore, we get the required straight line as y = 13.8482 — 0.6842x. When x = 10.5, from y = 13.8482 — 0.6842x y = 18.8482 — 0.6842 (10.5) 6.664119 Of a Second-Degree Parabolic Curve To fit a second-degree curve or parabola y = a+bx + cx Suppose a data consisting of a set of n pair of values Ga, yD, i= 1, 2...N. If x and y are two variables for this data, then to find the second-degree parabolic curve between them from given pair of values. We must find a, b and c such that it gives the best fit to the given data. We use principle of least squares to do this. The normal equations obtained by method of least squares are Sy = Na+b=x+e Ex? oxy = a txt b Ex? +e Ex Ex*y = a Dx? +b Ex8 +e Ext ... (5.7.2) Where N is the number of observations in a data and a, b and ¢ are constants to be determine. Solving above equations simultaneously for a, b and c we get the required quadratic curve.Procedure to fit a parabolic curve y = a + bx + cx? If x and y are given data, then we consider the following steps Write the normal equations for a second-degree | parabolic curve y = a + bx + cx”. Prepare a table with the columns x, y, xy, x2, x2y, | x3, x4, Complete the table with all entries in each column. |Find the sums 2 x, Dy, D xy, D x, d xy, =x", Ux! i.e. find sum of each column. Substitute the values of sums obtained in step 3 in normal Equations (5.7.2). Step 5 gives us three equations in three unknowns. Solve these equations simultaneously to get the values of constants a, b and c. Substitute the value of a, b and cobtained in step Gin y =a + bx + cx? to get a required parabolic curve. Ex. 5.7.6 : Fit a second-degree parabola to the following data x|/3.0/35/ 4 1/45] 5 |55] 6 3.113.3]3.6| 4 | 4.7] 54] 6.1Son.: Lety —a + bx + x be a second-degree parabolic "To At this second-degree parabolic curve, the normal equations are Sy = Na+bex +e Dx SExy = abe +b Ex? +e Sx Exty = a dx® +b Ex +6 Ext We construct a table to find the values of constants a, band & x ¥ xy — lps 3.0 | 3.1 | 9.30 | 9.00] 27.00 | 27.00 | 81.00 3s | ss | 11.55 |12.26| 40.425 | 42.875 |150.0625) a.o | 36 | 14.40 [16.00] 57.60 | 64.00 | 256.00 as | 4.0 | 18.00 |20.25| 51.00 | 91.125 |410.0625 s.o | 4.7 | 23.50 |26.00| 117.50 | 125.00 | 625.00 66 | 5.4 | 29.70 |30.25| 163.25 [166.375 015.0625 eo | 6.1 | 36.60 |s6.00| 219.60 | 216.00 | 1206.00 => Pax? <[ex?y = [exe = [ext 1.60 [30.20 [142.06 |148.75|707.a75 |vaz.a75 |s7za9.1676| Here N = 7.Substitute the values of the sums obtained from table in normal equations, ‘The normal equations become 30.20 Ta + 31.50b + 148.75c 143.05 = 31.5a + 148.75b + 732.375¢ 707.375 = 148.75a + 732.375b + 3733.1875¢ Solving above equations simultaneously for a, b and c we get a = 4.3929, b =— 1.1643, c = 0.2429 Put the values of a, b and c in y = a + bx + cx’ Therefore, we get the required parabola as y_= 4.3929 — 1.1648x + 0.2429x? 2 Ex. 5.7.7 : Fit a second-degree parabola to the following data aj2|{sl|4|s5|e6|]71]slo‘son. curve. To fit this second-degree parabolic curve, the normal equations are Let y a + bx + ex” be a second-degree parabolic Sy = Na+bex+e Ex Say = abx+b Sx* +e Sx3 Sty = a Ext 4b Ex3 46 Ext We construct a table to find the values of constants a, band c. x |» los | = [sy se 1 2 2 1 2 2 2 2 6 a2 4 2a 5 16 [3 7 | 2a 2 63 27 S81 4 s | s2 as | i258 64 256 5s | 10 | 50 | 25 | 250 | 125 625 s | 11 | 66 | se | s06 | 216 1296 z |a1 | 77 | 409 | ss0 | sas 2401 s | 10 | so | 64 | 640 | sia 4006 2 eo | si | si | 729 | 720 6561 zx [ey [exw [ee xt [= 4s |= 74] =421|= 255 = 15333Here N = 9. Substitute the values of the sums obtained from ble in normal equations, ‘The normal equations become 74 = 9a+45b + 285 421 = 45a + 285b + 2025c 2771 = 285a + 2025b + 15333¢ Solving above equations simultaneously for a, b and c we get a =— 0.9286, b = 3.5232, ¢ Put the values of a, b and c in y = a + bx + cx” ‘Therefore, we get the required parabola as — 0.9286 + 3.5232x — 0.2673x? — 0.2673 y= Ex. 5.7.8 : Using least square method fit a second- degree parabola y = a + bx + cx’ to the following data © fel ° 4 2 0.620 | 2.880 | 5.378 y | —3.150 | — 1.390 ae NUE Eero1. Fit a straight line for the following data xlels[s[s [7]eliol]a{o|7 yvisl7[7lio{s|slio}elsle 2. In a study between the amount of rainfall GQ and the quantity of air pollution (y) removed the following data were collected fit a straight line for the following data xj 43 | 4s [59 | 56 | 61 | 52 | 358 | 24 y | 126 | 12.1 | 116 | 11.8 | 11.4 | 11.8 | 13.2 | 14.4 3. Fit a straight line to the following data x | 41 | 51 | 61 | 71 | 81 | 94 y | 10 | 12 | 14 | 16 | 18 | 20 4. Fit a straight line of the form, y = a + bx to the following data and estimate the value of y for x = 3. xjo{l1|2ti{s3]4 yililis|sas}asles CECE Eve 5. Fit a straight line to the following data x | 100 [120 [140 [160 | 180 | 200 | y | 0.45 | 0.55 | 0.60 | 0.70 | 0.80 | 0.85y =4+0.5x —y = 15,5324 - 0.6842 x —y=1.8+0.2x —y = 0.72 + 1.338x — 0.0476 + 0.0041x6. Fit a parabola from the following data 7. Fit a parabola from the following data 1.0 [ 1.5 | 20] 25 [30 1.4 [43 | 16 | 20/27 oI) 8. Fit a straight line from the following data Zlele [= [ao] aa [ss [0 vi«[els2}se]20|2s} 20 | s2 9. Fit a parabola from the following data 3 s 6 4 44 27 39 19 10. Fit a second-degree curve from the following data @-[-]e=lelelele 2 9 | 25 | 34 | 40 | zo | 30 | 40 906. —y=2-0.5x + 0.2x" 7. —y = 1.0357 - 0.1929x + 0.2429x” 8 -y=2x 9. -y=16+0.2x+x" 10. —y =(7.6414 — 0.758x + 1.0836)"