Obtaining EQ-5D-3L utility index from the health status scale of traditional Chinese medicine (TCM-HSS) based on a mapping study

Table 3 Predictive performance evaluation of alternative models (N₂ = 213^#)

	MAE	RMSE	Rho	Rank MAE	Rank RMSE	Rank Rho	Final rank
OLS1	0.0627	0.1032	0.6354	15	10	15	12
OLS2	0.0638	0.1002	0.6351	16	6	21	14
OLS3	0.0606	0.1036	0.6454	13	11	10	9
OLS4	0.0613	0.1009	0.6468	14	7	9	7
Tobit1	0.2211	0.2511	0.6354	28	23	15	25
Tobit2	0.2181	0.2477	0.6367	25	21	13	23
Tobit3	0.2202	0.2514	0.6508	27	24	5	22
Tobit4	0.2189	0.2500	0.6555	26	22	1	18
CLAD1	0.0891	0.1218	0.6354	19	19	15	21
CLAD2	0.0861	0.1205	0.6373	18	17	12	17
CLAD3	0.0797	0.1157	0.6541	17	14	3	9
CLAD4	0.1531	0.1849	0.6470	24	20	8	19
GLM1	0.1075	0.3654	0.6354	20	25	15	24
GLM2	0.1264	0.4873	0.6313	23	28	22	27
GLM3	0.1196	0.4606	0.6283	21	26	23	26
GLM4	0.1231	0.4822	0.6235	22	27	24	27
TPM1	0.0585	0.1015	0.6354	12	8	15	11
TPM2	0.0563	0.0947	0.6377	6	2	11	3
TPM3	0.0571	0.1021	0.6479	10	9	7	6
TPM4	0.0554	0.0974	0.6550	4	3	2	1
ALDVMM1	0.0580	0.0987	0.6354	11	4	15	7
ALDVMM2	0.0570	0.0941	0.6361	9	1	14	4
ALDVMM3	0.0568	0.1037	0.6505	7	12	6	5
ALDVMM4	0.0554	0.0996	0.6517	5	5	4	2
OLOGIT1	0.0570	0.1196	0.4368	8	16	28	19
OLOGIT2	0.0535	0.1112	0.4720	3	13	27	14
OLOGIT3	0.0513	0.1213	0.4894	2	18	26	16
OLOGIT4	0.0499	0.1158	0.5363	1	15	25	13

OLS, ordinary least square; CLAD, Censored least absolute deviations; GLM, generalized linear model; ALDVMM, adjusted limited dependent variable mixture model; TPM, two-part model; OLOGIT, ordinal logistic regression; MAE, mean absolute error; RMSE, root mean squared error; Rho, Spearman's rank correlation coefficients;1–4, four model specifications
^#Validation set

ISSN: 1477-7525