The independent component analysis with the linear regression – predicting the energy costs of the public sector buildings in Croatia

Abstract

In the European Union, the public sector buildings are considered significant energy consumers and are, thus, the subject of several directives that aim to ensure the renovation of existing and the construction of new buildings as nearly zero-energy buildings. Therefore, as part of the decision making, it is necessary to properly plan the renovation or construction. This research provides models for predicting the energy costs of the public sector buildings, which are dependent upon its characteristics (i. e., constructional, occupational, energy, etc.). For this purpose, a real data set of Croatian public buildings was used, which included 150 variables and 1724 observations. Since the data set consisted of a large number of variables, the motivation for the dimensionality reduction was addressed first. Then, the independent component analysis, the principal component analysis, and the factor analysis were performed as the dimensionality reduction methods for variable extraction. The results of these analyses were used as inputs for modelling the energy costs of the public sector buildings. The obtained models were compared to the model built on original variables. The obtained models show the application potential in decision making for building renovation and construction in the public sector of Croatia, whereas the best performance of prediction in terms of RMSE and SMAPE was achieved by the model that integrated the independent component analysis with the linear regression.

Blind Source Separation with Strength Pareto Evolutionary Algorithm 2 (SPEA2) Using Discrete Wavelet Transform

“…In BSS, which is also a single-objective optimization method, the estimation of the source signals is provided by using the non-Gaussianity, sparsity and temporality properties of the signals. In single-objective optimization, while trying to reach a solution based on one of these features, the performance is increased by examining more than one feature with the use of Multi-Objective Optimization [6,7]. The performance of MOO methods used to separate mixed signals is often vital.…”