Daily surgery caseload prediction: towards improving operating theatre efficiency

Table 1 Daily surgery demand prediction results (MAPE)

	Emergency surgery			Elective surgery			Overall surgery
	1-year %	2-year %	3-year %	1-year %	2-year %	3-year %	1-year %	2-year %	3-year %
Rolling window	11.79	12.05	13.66	13.26	12.61	12.97	9.52	9.83	11.31
Regression (Linear)	N/A	13.71	13.79	N/A	17.79	15.84	N/A	13.49	12.85
Regression (Poisson)	N/A	14.17	14.71	N/A	18.90	16.68	N/A	13.99	13.77
Regression (Negative binomial)	N/A	14.13	14.88	N/A	19.12	16.67	N/A	14.08	13.86
Decision tree	15.69	17.97	18.30	14.55	13.91	14.72	15.99	13.97	12.14
Random forest	11.73	11.50	11.63	16.41	11.51	12.99	10.44	9.86	9.88
SVR (Linear)	19.13	16.50	17.50	15.61	15.62	15.19	35.81	37.93	36.96
SVR (RBF)	22.35	19.37	17.86	13.60	13.65	13.78	74.68	57.40	48.15
SVR (Sigmoid)	20.49	51.91	50.95	15.19	18.94	21.02	60.37	50.05	51.49
SVR (Poly)	44.12	26.67	23.63	19.23	21.31	15.55	90.15	48.84	36.43
Bagging regressor	12.86	11.97	12.50	13.62	11.90	14.25	10.62	10.42	10.26
Gradient boosting regressor	13.25	11.69	11.27	11.63	16.52	14.00	10.74	11.21	10.61
XGBoost regressor	13.60	13.65	14.89	37.18	15.74	40.92	15.06	11.64	11.22
Ensemble regressor	12.36	11.46	11.61	12.99	12.69	13.30	10.33	9.97	9.82

ISSN: 1472-6947