Machine Learning-Based CO₂ Prediction for Office Room: A Pilot Study

Table 2

Comparison of various models on the basis of statistical parameters.

S.N.	Methods	Statistical parameters for CO2							Standard deviation
				MSE	RMSE	MAE	MAPE	NS
1	Optimized GPR	0.98874	0.97761	17.64568	4.20068	3.35098	0.4325	0.9817	25.5432
2	GPR	0.98259	0.96548	27.08724	5.20454	4.25433	0.8393	0.9627	25.1313
3	Optimized EL	0.96447	0.93020	53.23267	7.29607	5.88308	1.1669	0.9267	24.4882
4	Optimized DT	0.95758	0.91696	60.30003	7.76531	6.00405	1.8131	0.9169	25.8812
5	DT	0.93714	0.87823	88.42094	9.40324	7.14055	1.4101	0.8782	25.3288
6	ANN	0.92064	0.84758	111.5761	10.56296	8.24138	1.7404	0.8314	25.7913
7	EL	0.89592	0.80267	156.4927	12.50970	9.97965	1.9695	0.7845	20.5966
8	LR	0.89566	0.80221	143.6322	11.98467	9.78463	1.9341	0.8022	24.2076
9	Optimized SVM	0.89349	0.79832	147.8662	12.16002	9.64578	1.9098	0.7964	24.5402
10	SVM	0.89044	0.79288	153.0833	12.37268	9.90023	1.9642	0.7892	24.1408