These data can be used to develop and validate statistical models,
such as the regression models considered in this paper. Regarding engineering models, the National
Renewable Energy Laboratory, in a continuous effort, has developed the software BEopt [56]. BEopt
is a no-cost software with a friendly user interface capable of evaluation of residential
building designs and identification of cost-optimal efficiency packages at various levels of
whole-house energy analysis. The data from smart meters will make it possible to calibrate models
developed with BEopt in an automatic manner using the Autotune methodology [57]. However, since
engineering software requires inputs on the characteristics of the dwelling that it may need the
participation of qualified personnel for accuracy. Therefore, the authors favor statistical models
because they can be developed without the need of input parameters that describe the building
characteristics. In order to illustrate the application of linear regres- sion approaches as
treated in Section 2, a case study to estimate the
total energy consumption is given in the following section.