Correlation and Regression Analysis

Similar documents
Chapter Two. An Introduction to Regression ( )

Statistics. Correlational. Dr. Ayman Eldeib. Simple Linear Regression and Correlation. SBE 304: Linear Regression & Correlation 1/3/2018

Multiple Choice Test. Chapter Adequacy of Models for Regression

Chapter Business Statistics: A First Course Fifth Edition. Learning Objectives. Correlation vs. Regression. In this chapter, you learn:

12.2 Estimating Model parameters Assumptions: ox and y are related according to the simple linear regression model

Objectives of Multiple Regression

b. There appears to be a positive relationship between X and Y; that is, as X increases, so does Y.

UNIT 7 RANK CORRELATION

Simple Linear Regression

Chapter 13 Student Lecture Notes 13-1

Homework Solution (#5)

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

Statistics MINITAB - Lab 5

Summary of the lecture in Biostatistics

Analyzing Two-Dimensional Data. Analyzing Two-Dimensional Data

STA302/1001-Fall 2008 Midterm Test October 21, 2008

Study of Correlation using Bayes Approach under bivariate Distributions

Simple Linear Regression and Correlation.

The Mathematics of Portfolio Theory

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions.


UNIT 6 CORRELATION COEFFICIENT

( ) = ( ) ( ) Chapter 13 Asymptotic Theory and Stochastic Regressors. Stochastic regressors model

Linear Regression with One Regressor

Mean is only appropriate for interval or ratio scales, not ordinal or nominal.

CHAPTER 8 REGRESSION AND CORRELATION

Can we take the Mysticism Out of the Pearson Coefficient of Linear Correlation?

Example: Multiple linear regression. Least squares regression. Repetition: Simple linear regression. Tron Anders Moger

Probability and. Lecture 13: and Correlation

Correlation and Simple Linear Regression

ε. Therefore, the estimate

Comparison of Dual to Ratio-Cum-Product Estimators of Population Mean

Simple Linear Regression

ESS Line Fitting

Multiple Linear Regression Analysis

Summarizing Bivariate Data. Correlation. Scatter Plot. Pearson s Sample Correlation. Summarizing Bivariate Data SBD - 1

STATISTICAL PROPERTIES OF LEAST SQUARES ESTIMATORS. x, where. = y - ˆ " 1

Line Fitting and Regression

ECONOMETRIC THEORY. MODULE VIII Lecture - 26 Heteroskedasticity

Chapter Statistics Background of Regression Analysis

MEASURES OF DISPERSION

"It is the mark of a truly intelligent person to be moved by statistics." George Bernard Shaw

Chapter 4 Multiple Random Variables

Third handout: On the Gini Index

Multiple Regression. More than 2 variables! Grade on Final. Multiple Regression 11/21/2012. Exam 2 Grades. Exam 2 Re-grades

THE ROYAL STATISTICAL SOCIETY HIGHER CERTIFICATE

STA 108 Applied Linear Models: Regression Analysis Spring Solution for Homework #1

Lecture 1: Introduction to Regression

Functions of Random Variables

Midterm Exam 1, section 1 (Solution) Thursday, February hour, 15 minutes

Simple Linear Regression - Scalar Form

i 2 σ ) i = 1,2,...,n , and = 3.01 = 4.01

Regression and the LMS Algorithm

Lecture 3. Sampling, sampling distributions, and parameter estimation

Lecture 1: Introduction to Regression

Transforming Numerical Methods Education for the STEM Undergraduate Torque (N-m)

ρ < 1 be five real numbers. The

CHAPTER VI Statistical Analysis of Experimental Data

Midterm Exam 1, section 2 (Solution) Thursday, February hour, 15 minutes

9.1 Introduction to the probit and logit models

2SLS Estimates ECON In this case, begin with the assumption that E[ i

Lecture 8: Linear Regression

Module 7: Probability and Statistics

Chapter 2 Supplemental Text Material

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses

Johns Hopkins University Department of Biostatistics Math Review for Introductory Courses

Previous lecture. Lecture 8. Learning outcomes of this lecture. Today. Statistical test and Scales of measurement. Correlation

Module 7. Lecture 7: Statistical parameter estimation

L5 Polynomial / Spline Curves

CSE 5526: Introduction to Neural Networks Linear Regression

hp calculators HP 30S Statistics Averages and Standard Deviations Average and Standard Deviation Practice Finding Averages and Standard Deviations

University of Belgrade. Faculty of Mathematics. Master thesis Regression and Correlation

Chapter 14 Logistic Regression Models

Lecture Notes 2. The ability to manipulate matrices is critical in economics.

Econometric Methods. Review of Estimation

Lecture Notes Forecasting the process of estimating or predicting unknown situations

Sampling Theory MODULE V LECTURE - 14 RATIO AND PRODUCT METHODS OF ESTIMATION

Lecture 3 Probability review (cont d)

PGE 310: Formulation and Solution in Geosystems Engineering. Dr. Balhoff. Interpolation

Lecture 12 APPROXIMATION OF FIRST ORDER DERIVATIVES

Statistics: Unlocking the Power of Data Lock 5

Linear Regression. Can height information be used to predict weight of an individual? How long should you wait till next eruption?

CLASS NOTES. for. PBAF 528: Quantitative Methods II SPRING Instructor: Jean Swanson. Daniel J. Evans School of Public Affairs

4. Standard Regression Model and Spatial Dependence Tests

Algorithms behind the Correlation Setting Window

1 Lyapunov Stability Theory

arxiv:math/ v1 [math.gm] 8 Dec 2005

UNIT 2 SOLUTION OF ALGEBRAIC AND TRANSCENDENTAL EQUATIONS

Lecture Notes Types of economic variables

CS 2750 Machine Learning. Lecture 8. Linear regression. CS 2750 Machine Learning. Linear regression. is a linear combination of input components x

Linear Regression. Hsiao-Lung Chan Dept Electrical Engineering Chang Gung University, Taiwan

ENGI 4421 Joint Probability Distributions Page Joint Probability Distributions [Navidi sections 2.5 and 2.6; Devore sections

r y Simple Linear Regression How To Study Relation Between Two Quantitative Variables? Scatter Plot Pearson s Sample Correlation Correlation

CHAPTER 2. = y ˆ β x (.1022) So we can write

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

best estimate (mean) for X uncertainty or error in the measurement (systematic, random or statistical) best

Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class)

Generalized Minimum Perpendicular Distance Square Method of Estimation

Physics 114 Exam 2 Fall Name:

Maximum Likelihood Estimation

Transcription:

Chapter V Correlato ad Regresso Aalss R. 5.. So far we have cosdered ol uvarate dstrbutos. Ma a tme, however, we come across problems whch volve two or more varables. Ths wll be the subject matter of the curret chapter. D. 5.. (Correlato) The correlato meas the stud of estece, magtude ad drecto of the relato betwee two or more varables. R. 5.. (Tpes of Correlato) We dstgush. Postve ad egatve correlatos If two varables chage the same drecto, the ths s called a postve correlato. (For eample: prce ad suppl) If two varables chage the opposte drecto, the the correlato s called a egatve Correlato. (For eample: prce ad demad). Lear ad o-lear correlatos. R. 5. 3. (Degree of Correlato) Degrees Postve Negatve Absece of correlato Perfect correlato + -,.75 Hgh degree ].75, [ ] [ Moderate degree ].5,.75 [ ]-.75, -.5[ Low degree ],.5[ ].5, [ R. 5. 4. (Methods of Determg Correlato) We shall cosder the followg most commol used methods: () Scatter Plot () Karl Pearso s coeffcet of correlato (3) Spearma s rak correlato coeffcet.

R. 5. 5. (Scatter Plot or Dot Dagram) I ths method the values of the two varables are plotted o a graph paper. Oe s take alog the horzotal ( ) as ad the other alog the vertcal ( ) as. B plottg the data, we get pots (dots) o the graph whch are geerall scattered ad hece the ame scatter plot. The maer whch these pots are scattered, suggest the degree ad the drecto of correlato. Let the degree of correlato be deoted br. Its drecto s gve b the sgs postve ad egatve. ) If all pots le o a rsg straght le the correlato s perfectl postve ad r =+ ) If all pots le o a fallg straght le the correlato s perfectl egatve ad r = ) If the pots le a arrow strp, rsg upwards, the correlato s hgh degree of postve. v) If the pots le a arrow strp, fallg dowwards, the correlato s hgh degree of egatve. v) If the pots are spread wdel over a broad strp, rsg upwards, the correlato s low degree postve. v) If the pots are spread wdel over a broad strp, fallg dowwards, the correlato s low degree egatve. v) If the pots are spread (scattered) wthout a specfc patter, the correlato s abset,.e. r = Though ths method s smple ad gves a rough dea about the estece ad the degree of correlato, t s ot relable. As t s ot a eact mathematcal method, t caot measure the degree of correlato. D. 5.. (Karl Pearso s Coeffcet of Correlato) Let X ad Y be two varates. = r : =. = = E. 5.. The followg table shows the aual proft [Mo ] ad aual costs of leasg computer equpmet [ ] of 5 frms:

( ) ( ) ( ) ( ) - 3-7 34 4 89 5-5 3-7 55 5 89 3 5-5 - 5 5 4-5 -5 5 5 5 - - 6 5-5 8-6 5 44 7 3 5-5 5 8 3-9 3 5 5 5 35 5 8 - - 5 4 35 5 33 3 65 5 69 4 3 45 5 4 3 5 4 4 5 5 3 6 4 9 5 5 6 4 8 4 6 Sum 45 3 8 5 457 45 3 = = 3, = = 5 5 8 r : =.88. 5 457 3

D. 5. 3. (Spearma s Rak Coeffcet of Correlato) Let R, =,,..., : raks of the characterstc X ' R, =,,..., : raks of the characterstc Y, 6 = ρ : = ' ( R R) ( ) ( + ) : E. 5.. The followg table shows the advertsemet costs (Y ) ad the reveues (X ) of a frm: Advertsemet Costs Reveue.4.8.9 4.4 4.8 3 3. 4 3.6 4 4. 48 Fd ad terpret the Spearma s rak coeffcet of correlato. Soluto: Advertsemet Costs Reveue R ' R ( R R ).4...8...9 4 3 3.5.5.4 4 4 3.5.5.8 3 5 5.. 3. 4 6 6.. 3.6 4 7 7.. 4. 48 8 8...5 8 ' = 8, ( R R) =.5, = 6.5 ρ : =.994. 7.8.9 There s therefore a strog degree of correlato betwee X ady. ' 4

D. 5. 4. (Regresso Fucto) A regresso fucto descrbes the relatoshp betwee depedet varable Y ad (at least oe) depedet or eplaator varable X : * = f ( ) R. 5. 6. (Least Square Method) The coeffcets of a regresso fucto ca be estmated b usg the Least Square Method: = ( ) S(...) = * M! R. 5. 7. (Smple Lear Regresso) The coeffcets of a smple lear regresso fucto * = a + a ca be obtaed b solvg the followg sstem of lear equatos: a a + a + a = = = = = = = D. 5. 5. (Coeffcets of Correlato ad Determato). The coeffcet of correlato of a lear regresso fucto s defed b = = = r : =, = = = = = = r +. r ( r ) s called coeffcet of determato. E. 5. 3. I a stud of how the productvt 4 frms depeds o the degree of automato, the followg data have bee made avalable : The term regresso was frst used b Sr Fracs Galto (8-9), who studed the relatoshp betwee heghts of chldre ad heghts of ther parets. 5

Frm Productvt Degree of Automato ( %) 3 4 3 3 8 36 4 3 4 5 3 4 6 33 47 7 34 56 8 37 54 9 38 6 4 55 4 6 43 67 3 45 69 4 48 76. Plot a scatter dagram.. Fd a approprate regresso fucto. 3. Calculate the coeffcets of correlato ad determato ad terpret them. 4. Predct the level of productvt for a automato degree of 8%. Soluto:. 5 45 4 35 3 5 5 3 4 5 6 7 8 6

. 3 64 4 4 4 3 7 9 576 3 8 36 8 96 784 4 3 4 6 9 5 3 4 7 68 96 6 33 47 55 9 89 7 34 56 94 336 56 8 37 54 998 96 369 9 38 6 8 36 444 4 55 35 6 4 6 5 37 68 43 67 88 4489 849 3 45 69 35 476 5 4 48 76 3648 5776 34 total 49 74 697 434 838 4a + 74a 74a + 434a = = 49 697 The soluto of the above sstems elds: a = 7.356, 5435 a =.. Therefore, we have the regresso fucto: * = 7.356+.5435 6 5 4 3 4 6 8 7

3. 4 697 74 49 r=.9687 ( 4 434 74 74) ( 4 834 49 49) r =.9384 Because of r > the productvt s drectl depedet o the degree of automato. Chages the productvt are up to 94% due to chages the degree of automato. 3. *( 8) = 7.356+.5435 8= 5 D. 5. 6. (Tred Fucto) A tred fucto s a specal case of a regresso fucto where the depedet varable s the factor tme. D. 5. 7 (Coeffcet of Determato) The coeffcet of determato s defed as: r : = SSR SST where: = * SSR= ( ) : sum of square due to regresso = SST = ( ) : total sum of squares R. 5. 8. The formula gve for the coeffcet of correlato gve D. 5. 5. s a specal case of the formula D. 5. 7. for lear regresso. E. 5. 4. The followg table shows the cosumpto per head of butter ( kg) a certa coutr the ears 998-4: Year 998 999 3 4 Cosumpto per head 9. 9.5 9.7 9.6 9.9.5.8. Ft a lear tred b regresso aalss.. Predct the cosumpto per head of butter the ear 5. 8

Soluto. ear 998-3 9. 9-7.6 999-9.5 4-9. - 9.7-9.7 9.6. 9.9 9.9 3.5 4. 4 3.8 9 3.4 69. 8 7 a = 69. 9.886 7 =, 7 a. 8 5 = =. We, therefore, have the tred fucto:. * =.885+.5. *( 4) =.885+.5 4=.9 E. 5. 4. (Learsato of a No-Lear Fucto) The followg table shows the umber of cattle a certa coutr the ears 99-4: Year 99 994 995 998 4 No. of cattle (( 5 ). 3.3 4.6 9.6 3.7 33.. Ft a tred b the regresso fucto * = a a.. Predct the umber of cattle the ear 5. 3. Calculate ad terpret the coeffcet of determato for the regresso. Soluto: Let * = a a lg * = lga + lga lg * = lga + lga. Y*: = lg *, A = lga, A = lga. The we have: 9

Y* = A + A A = + A A + A 6A + 37A 37A + 35A = Y = = = = Y = ear lg lg 99..439.439 994 3 3.3.385 9 3.3755 995 4 4.6.6435 6 4.6574 998 7 9.6.96 49 9.458 9 3.7.37475 8.3775 4 3 33..5983 69 9.75779 37 5.3 7.5643 35 5.467 = = 7.5 5.5 The soluto of the above sstems elds:. e. A =.645439, 43443 A =., a =.49777, 965659 a =. Therefore, we have the regresso fucto: * =.5.97 6 5 4 3 5 5. 4 5 *(4) =.5.97 = 37. ( ) (Last revsed: 7..)