Research on the Combined Prediction Model of Residential Building Energy Consumption Based on Random Forest and BP Neural Network

Table 2

Description of the data set.

Variables	Property name	Abbreviations	Number of possible values	Minimum	Maximum	Average	Median	Standard deviation
X1	Relative compactness	RC	12	0.62	1.00	0.76	0.76	0.10
X2	Surface area	SA	12	2.00	808.50	670.83	661.50	91.28
X3	Wall area	WA	7	3.00	416.50	318.08	318.50	45.05
X4	Roof area	RA	4	4.00	220.50	176.37	147.00	45.56
X5	Overall height	OH	2	3.50	7.00	5.24	5.00	1.75
X6	Orientation	OR	4	2.00	6.00	3.50	4.00	1.12
X7	Glazing area	GA	4	0.00	7.00	0.24	0.25	0.27
X8	Glazing area distribution	GAD	6	0.00	8.00	2.81	3.00	1.56
Y1	Heating load	HL	583	1.00	43.10	22.27	18.89	10.11
Y2	Cooling load	CL	636	2.00	48.03	24.55	22.07	9.54