Development of Regression Models considering Time-Lag and Aerosols for Predicting Heating Loads in Buildings

Table 16

Residual statistics for a DOE medium office building depending on different models.

Residual statistics^a
		Minimum	Maximum	Mean	Std. deviation	N
Modified model^b	Predicted value	−41989976.00	47291244.00	5299326.20	8723871.147	20,350
	Residual	−34259568.000	135437520.000	0.000	12626073.134	20,350
	Std. predicted value	−5.421	4.813	0.000	1.000	20,350
	Std. residual	−2.713	10.724	0.000	1.000	20,350
New model^c	Predicted value	−5426438.00	19135408.00	1919966.69	3352476.356	4807
	Residual	−19135408.000	73064632.000	0.000	6504633.271	4807
	Std. predicted value	−2.191	5.135	0.000	1.000	4807
	Std. residual	−2.939	11.223	0.000	0.999	4807
Limited model^d	Predicted value	−15162106.00	27670688.00	5299326.20	7446224.483	20,350
	Residual	−22803048.000	150226688.000	0.000	13419291.767	20,350
	Std. predicted value	−2.748	3.004	0.000	1.000	20,350
	Std. residual	−1.699	11.193	0.000	1.000	20,350

^aDependent variable: heating load of a DOE medium office building. ^bPredictors: (constant), visibility, diffuse radiation, atmospheric station pressure, wind speed, total sky cover, dew point temperatures, direct normal radiation, relative humidity, global horizontal radiation, and horizontal infrared radiation; ^cpredictors: (constant), precipitable water, direct normal radiation, total sky cover, AOD, atmospheric station pressure, dew point temperatures, dry bulb temperatures, horizontal infrared radiation; ^dpredictors: (constant), total sky cover, wind speed, dry bulb temperatures, global horizontal radiation, relative humidity, atmospheric station pressure, dew point temperatures.