01.03.2013 Views

Applied Statistics Using SPSS, STATISTICA, MATLAB and R

Applied Statistics Using SPSS, STATISTICA, MATLAB and R

Applied Statistics Using SPSS, STATISTICA, MATLAB and R

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

7.1 Simple Linear Regression 275<br />

Figure 7.3. Table obtained with <strong>STATISTICA</strong> containing the results of the simple<br />

linear regression for the Example 7.1.<br />

The value of Beta, mentioned in Figure 7.3, is related to the so-called<br />

st<strong>and</strong>ardised regression model:<br />

*<br />

Y = β x + ε . 7.6<br />

*<br />

i<br />

1<br />

*<br />

i<br />

i<br />

In equation 7.6 only one parameter is used, since Y <strong>and</strong><br />

variables (mean = 0, st<strong>and</strong>ard deviation = 1) of the observed <strong>and</strong> predictor<br />

variables, respectively. (By equation 7.5, β 0 = E[ Y ] − β1x<br />

implies ( Yi − E[<br />

Y ] ) / σ Y<br />

*<br />

*<br />

= β 1 ( xi − x)<br />

/ s X + ε i .)<br />

It can be shown that:<br />

*<br />

i<br />

*<br />

x i are st<strong>and</strong>ardised<br />

σ *<br />

β1 ⎟β1 ⎟<br />

⎛ Y ⎞<br />

=<br />

⎜ . 7.7<br />

⎝ s X ⎠<br />

The st<strong>and</strong>ardised *<br />

β 1 is the so-called beta coefficient, which has the point<br />

*<br />

estimate value b 1 = 0.98 in the table shown in Figure 7.3.<br />

Figure 7.3 also mentions the values of R, R 2<br />

<strong>and</strong> Adjusted R 2<br />

. These are<br />

measures of association useful to assess the goodness of fit of the model. In order<br />

to underst<strong>and</strong> their meanings we start with the estimation of the error variance, by<br />

computing the error sum of squares or residual sum of squares (SSE) 1<br />

, i.e. the<br />

quantity E in equation 7.2, as follows:<br />

∑<br />

2<br />

∑<br />

2<br />

SSE = ( y i − yˆ<br />

i ) = ei<br />

. 7.8<br />

Note that the deviations are referred to each predicted value; therefore, SSE has<br />

n − 2 degrees of freedom since two degrees of freedom are lost: b0 <strong>and</strong> b1. The<br />

following quantities can also be computed:<br />

1<br />

–<br />

SSE<br />

Mean square error: MSE = .<br />

n − 2<br />

– Root mean square error, or st<strong>and</strong>ard error: RMS = MSE .<br />

Note the analogy of SSE <strong>and</strong> SST with the corresponding ANOVA sums of squares,<br />

formulas 4.25b <strong>and</strong> 4.22, respectively.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!