Scribd
Upload
14 views8 pages

Correlation BVRamana

Uploaded by

Sahil Ghuge
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
14 views8 pages

Correlation BVRamana

Uploaded by

Sahil Ghuge
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 8
Curve Fitting, Regression and Correlation Analysis Pe yr=ODUCTION Approximating curve is the graph of data obtained though measurement or observation, Curve fitting site process of finding the “best fit” curve since dierent approximating curves can be obtained for tesamedata. Least squares method is the best curve fingmethod and is easily implemented on comput- esihan the other methods like method of moments, ehod of group averages, graphical method. We also consider curve fitting by axveighted and non-linear weighted least squares ‘approximation. Inthe unvariate case, a single variable say the Height of an Indian, is analyzed. Whereas in the bi- Yafatecase, two "numerical" variables are measured resulting in a pair of measurements for each m ber say the height and weight of an Indian; the age 424 the blood pressure of a person; amount spent ‘advertising and volume of sales; intake of nutri- {sfood and 1.Q of a student etc. In the correlation ‘alysis, one wish to find whether @ (mathematical) onship exists and measure the strength of such ip. In the regression analysis, the exact na fia form of mathematical equation (of the re is obtaned. While the correlation coefficient the “closeness”, the "regression equation for prediction (or estimation). sum of exponentials, lin- FITTING of finding equation ©! set of data. On the basis fa curve that of this mathematical equation, predictions can be made in many statist Relationship Relationship (or as variables may exist. ion) between two (or more) Examples: Blood pressure and age, rainfall and crop yield, volume of a cube and length of its side, consumption of food and weight gain, intake of drug and heart rate, height and weight, income and medi- cal care, nutrition and 1.Q. Scatter Diagram To find a mathematical relationship (equation) be- tween say two variables X and Y, plot the set of given N paired observations of X and Y ie. (X1, Yi), (Xa, Ya), «++ Xn ¥w) in the XY-plane. The resulting set of points is known as a scatter dia- gram. ‘Approximating Curve It isa smooth curve that approximates the given set of N data points plotted in the scatter diagram. Collocational polynomial For unequally spaced X's N coefficients doy, @2,---,aw-=1 in the pes Jocational polynomial Y(X) = 9 + aX + aX? ++ Lona can be determined so thatthe given set of NV data points (X1+ Yy)e-+- (Xe Yn) lies (collocates) on the curve (satisfies ‘the above equation) 30.1 Parabola oF quadratic ¢ 3. Yea taix4.., nth degree (polyno, Lagrange’ interpolation formula is uses ee Serre rtes pons Hs age and ‘the empirical or observed data contains errors (i.¢.. § ae random variables). In such exponential Bempe dee etacin Ie ya axe | pnts Y = f(x) containing few unknown parameters. Ber siy ren ‘Best fitting curve by method of least squares Let (6), Ys (Xa, Yad (ips Py) be a given set of NV data points. Let dj = ¥; ~ ¥; denotes the difference ‘between ¥; and the Curve Fiting by Least Squares () Least squares si Lati-o Dura yd ° Pina D xa De tale oF 0% = a 0X) =0 w ” te, O% = Nata OX linear and can be solved See: Normal equations of the mh degree Ext Bic teote ly by molipying @) O both sides 1X" and summing orm o> a ar =200; -a © rx -e Duta Le . ‘Thus the two unknown parameters 4° ee vit (Extrapolation) isto nd Y cor. toa value of X included between "0% io to the) given values of X- © Drenm ta DK : ern asx * Result 1; Psp PHC He aay Sohation: Lat Prater o ‘esheets ine LS.) of 0 Xs ermal Lr=swsaDx stag 9a, EE BE « Y= apt ak @ ae through (X. 7). Xatthr ° ivding by its normal equation, we get eee EX yin et tawny SoLSL passes through ¥. 7). Result 2: Prove that LSI. ¥ on X can be ex presses over igh HP), Pasta Seine y-F=ax-H e year isco stow tas = $33 Lr- seen le pxreaDreare wyxr-Ex5r any -0.12 at 001 level of significance. Hint: N = 6,30 X = 42.0 Y = 78.07 = 364, DXY = 48.6, 5 YF = 10.68 Sex = $00, Sry = 3.24, Sur = ~36.5¢ = 0.08017. Ans, a. Y(X) = 19 — 0.086X, ¥(5) = L47 1 confidence interval for a (1.6933, 2.1067) 1 = 3.58, reject oul hypothesis =33, Ex i355, bb Determine a 95% confidence interval for a. : Testibe hypothesis B = LOagainst < 1.0. 4. Testthe hyposhesisthata = Oagainstar # 0 at 005 level of significance. Ans, 2. Y =38296 + 0.9036X, ¥(25) = 26.4196 08012

You might also like