A comparative study on the prediction of the BP artificial neural network model and the ARIMA model in the incidence of AIDS

Table 5 MSE of 11 BP-ANN algorithms base on 3–12 neurons in the hidden layer

Algorithm	Number of neurons in the hidden layer
Algorithm	3	4	5	6	7	8	9	10	11	12
Traingd	0.597710	0.633182	0.566311	0.888439	0.778596	0.895304	1.025611	1.057920	0.425543	0.382488
Traingdm	0.003257	0.002775	0.003120	0.003124	0.003389	0.003088	0.003015	0.003237	0.003293	0.003116
Traingda	0.002978	0.002820	0.003169	0.002910	0.002736	0.003304	0.002894	0.003054	0.003250	0.002987
Traingdx	0.004025	0.003410	0.003930	0.003967	0.003496	0.002735	0.003296	0.003464	0.003186	0.003055
Trainrp	0.004357	0.004044	0.004410	0.004013	0.004315	0.004017	0.004347	0.004304	0.004002	0.003873
Traincgf	0.004123	0.004409	0.003290	0.003908	0.003490	0.004200	0.004084	0.003001	0.004252	0.004482
Traincgp	0.003626	0.004292	0.003758	0.002979	0.003060	0.003433	0.004048	0.004186	0.004122	0.003273
Traincgb	0.003661	0.002862	0.002901	0.002945	0.003922	0.003591	0.003041	0.003591	0.002966	0.002799
Trainscg	0.004381	0.004148	0.004444	0.004257	0.004166	0.004352	0.004403	0.004491	0.003700	0.004392
Trainoss	0.003074	0.003489	0.002980	0.002927	0.003281	0.002651	0.002948	0.003391	0.003032	0.003330
Trainlm	0.002369	0.002293	0.002123	0.001863	0.002042	0.002313	0.002330	0.002365	0.002445	0.002491

ISSN: 1472-6947