基于熵权法结合层次分析法和反向传播神经网络优选乌梅肉炮制工艺OA

Objective To optimize the best processing technology of Wumei(Mume Fructus,MF)flesh and determine the best processing parameters.Methods The contents of 5-hydroxymethylfurfural(5-HMF),neochlorogenic acid,chlorogenic acid,cryptochlorogenic acid and caffeic acid were determined by HPLC.The content of polysaccharide in MF flesh was determined by ultraviolet-visible spectrophotometer.The content of each chemical component,the content of water-soluble extracts and the appearance traits were selected as the evaluation indexes,and the weight coefficient of each evaluation index was determined by analytic hierarchy process(AHP)-entropy weight method hybrid weighting method.The L9(34)orthogonal test was carried out with water addition,moistening time and steaming time as factors.Back propagation neural network(BPNN)was used to predict and verify the optimal processing parameters of MF flesh.Results The BPNN optimization process was superior to the orthogonal screening process,and the prediction results were stable and reliable.The composite score was the highest,which was 83.26.The best steaming process of MF flesh was 22.5 kg of water per 100 kg of MF flesh,moistening for 60 min,and steaming for 30 min.Conclusion The determined processing technology of MF flesh is scientific and feasible,which can provide scientific data support for the industrial production of decoction pieces,and also provide a theoretical basis for ensuring the quality and clinical application of MF flesh.