| Xinyang Maojian tea industry has become an important pillar of economic development and poverty alleviation in Xinyang,The problems in the development of the Xinyang Maojian tea industry include the lack of marketing strategy and the unpredictability of production.Therefore there are two key needs for the development of the Xinyang Maojian tea industry:scientific and intelligent production forecasting tools and an evaluation system of marketing strategy.This paper proposes a prediction model of Xinyang Maojian yield based on PSO-BPNN and multi-source data fusion.Firstly,in terms of the selection of model data,it emphasizes the fusion of multi-source data,and the collection contents include the remote sensing image data of Resources No.1 satellite,meteorological data,production data of Xinyang Maojian,basic map data and Beidou positioning data,etc.This paper creatively uses the normalized differential vegetation index as the intermediate variable.The experiment shows that there is a strong linear relationship between the difference in normalized vegetation index and the mean temperature.Therefore the fusion of meteorological data and normalized vegetation index data can achieve high-performance yield prediction.In terms of model selection,the input-output of this study has a high nonlinear relationship,so the prediction method adopts back-propagation neural network(BPNN),which can solve the solid nonlinear input-output relationship.In order to speed up the training of BPNN model,particle swarm optimization(PSO)is adopted in this paper to optimize the initial weight and threshold of BPNN model.At the same time,in order to effectively overcome the shortcomings of slow convergence speed and easy to fall into local optimal of PSO algorithm,dynamic reduction of inertia weight is carried out in this paper to enhance the global search ability of PSO algorithm,and self-mutation operator is introduced.Further reduce the probability of the algorithm falling into local optimal.This paper also puts forward the index system of Xinyang Maojian’s marketing strategy based on the Analytic Hierarchy Process(AHP)and Fuzzy Comprehensive Evaluation(FCE).The combination of AHP and FCE can effectively determine the best decision.Based on the Delphi method,the evaluation index system of Xinyang Maojian tea marketing strategy based on the subjective experience of experts was given.Based on the AHP+FCE method,the weights and scores of each index are given,which has an important reference role for the formulation and evaluation of Xinyang Maojian’s marketing strategy. |
