Part 13- Simple linear regression - The University of Jordan
Part 13- Simple linear regression - The University of Jordan
Part 13- Simple linear regression - The University of Jordan
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>University</strong> <strong>of</strong> <strong>Jordan</strong> Agricultural Statistic (605150)Faculty <strong>of</strong> AgricultureDr. Amer SalmanDept. <strong>of</strong> Agri. Econ. & Agribusiness22222 2( Y − Y ) = ( Yˆ− Y ) + ( Y − Yˆ) ,[ y = yˆe ]∑ i ∑ i i ∑ i i ∑ i ∑ i+ ∑Total variation = Explained variation + Residual variationIn Y (or total in Y (or <strong>regression</strong> in Y (or errorSum <strong>of</strong> squares) sum <strong>of</strong> squares) sum <strong>of</strong> squares)TSS = RSS + ESSDividing both sides by TSS:RSS ESS1 = +TSS TSS2<strong>The</strong> coefficient <strong>of</strong> determination or R is then defined as the proportion <strong>of</strong> the totalvariation in Y explained by the <strong>regression</strong> <strong>of</strong> Y on X.iR2RSS= = 1−TSSESSTSS2R Can be calculated by:R2∑∑∑∑22= yˆiei= 1−2y y2ii∑∑Where: y ˆ = ( − ) 2iYiYi22∑ei= ∑ ( Y i− Yˆi)22∑ yi= ∑ ( Yi− Y )2 ˆ2R Ranges in value from 0 (zero when the estimated <strong>regression</strong> equation explains non<strong>of</strong> the variation in Y) to 1 (when all points lie on the <strong>regression</strong> line).<strong>The</strong> Correlation Coefficient:r<strong>The</strong> correlation coefficient is given by:∑ x2 iyi∑== ˆxiyiRb12 2x y yi=2∑i∑i∑- 14 -