Application of a Model Based on Rough Set Theory (RST) to Estimate the Energy Efficiency of Public Buildings

Abstract

The study was carried out on a group of 85 public buildings, which differed in type of use, construction technology and heating systems. From the collected data, a set of qualitative and quantitative variables characterizing them in terms of heat demand was extracted. In this paper, the authors undertook to test the suitability of a model based on rough set theory (RST), which allows the analysis of imprecise, general and uncertain data. To obtain input data for the RST model in quantitative form, the authors used an alternative approach, which is a method based on the thermal properties of buildings. The quality of the predictive model was evaluated based on the following indicators, such as the coefficient of determination (R2), the mean bias error (MBE), the coefficient of variance of the root mean square error (CV RMSE) and the mean absolute percentage error (MAPE), which are accepted as statistical calibration standards by ASHRAE (American Society of Heating, Refrigerating and Air-Conditioning Engineers). A quality-acceptable predictive model must meet the calibration conditions: MBE ±5%, CV RMSE < 15% and R2 > 0.75. For the analyzed RST model, the following values of evaluation indicators were obtained: MBE = −1.1%, CV RMSE = 11.8% and R2 = 0.91. The evaluation results obtained gave rise to the conclusion that the method used, which is based on a limited amount of data describing buildings, gives good results in estimating the unit rate of energy demand for heating.