The analysis of thermal end energy consumption results in a model with an explanatory percentage of variance of R2 = 53%.
The calculated coefficient "bi" of the best fit model to predict the heating energy consumption (y1) of new schools on the basis of the
regression model gives the final regression equation with the independent variables "energy standard" (x2), "type of use" (x3) and
"energy source" (x6) (see Table 3). The following F-test shows a
very high significance, whereas adding of a further independent
variable will stop the forward multiple regression as the limit value of significance (P = 5.0%) will be exceeded and the model with more independent variables will not explain a higher proportion of variance. Therefore the variables "heated gross area" (x5), "year of construction" (x1) and "canteen kitchen" (x4) were excluded