Hence, the non-parametric Wilcoxon test was conducted to test if the “mean differences” between predicted energy use and actual energy use were statistically significant. The differences were also statistically tested using the parametric paired t-test. Results were similar to that of Wilcoxon test. Results of paired t-test were more convenient to explain and therefore, the same was presented in Table 4 (in Section 3). The level of significance (sig.