[Retracted] Sales Forecast of Marketing Brand Based on BP Neural Network Model

Table 2

Comparison of prediction results and errors of the two methods.

Year	Actual value	Regression analysis			Neural network method
		Predictive value	Absolute error	Relative error	Predictive value	Absolute error	Relative error
2017	2684.7	2150.9	-533.8	−19.88%	2731.5	46.8	1.74%
2018	2691.4	2746.8	55.4	2.06%	2768.1	76.7	2.85%
2019	2867.8	4046.5	1178.7	41.10%	2860.1	−7.7	−0.27%
2020	3032.9	5499.6	2466.7	81.33%	3011.5	−21.4	−0.71%
Average value	2819.2	3610.95	791.75	26.15%	2842.8	23.6	0.90%

Note. Absolute error = predicted value − actual value; relative error = absolute error/actual value.