Determination of Regression Parameters for the Thermal and Energy Components of Smart Homes

Abstract

Determination of the regression parameters of different data sets which consist of the thermal and energy components of smart homes is carried out. Analyzed data were automatically gathered by installed sensors under UK REFIT Smart Home project and published for an open access. In order to carry through the statistical analysis of the given data, the research stages have been divided into three steps. Firstly, the given data have been stored into a specially developed sqlite database. Secondly, the data prepossessing (cleaning, smoothing and resampling) and visualizing have been provided using Python libraries (pandas, seaborn, matplotlib). Then, the cleaned and smoothed data have been used for the statistical regression analysis within which different hypotheses of relations have been investigated. Conducted study of the thermal and energy components of the data has shown a suitability of such data for a forecasting and formulation of the inverse problem of determination of the smart home thermophysical parameters.