Research On Cold Chain Logistics Distribution Of Rural Agricultural Products In Changsha County Based On Machine Learning

China’s agricultural cold chain logistics has broad prospects for development.Fresh e-commerce has developed rapidly with huge market space,and has received the policy support of the national government.In order to speed up the development of cold chain logistics of rural agricultural products in the county and improve the logistics operation efficiency,combined with the characteristics of cold chain logistics of agricultural products and rural distribution in the county,based on the machine learning method,Taking Changsha County,Hunan Province as the research area,the research on cold chain logistics distribution of rural agricultural products in the county is carried out.The main work is as follows:(1)Combined with the current situation and characteristics of agricultural cold chain logistics and county-level rural three-level distribution,this paper analyzes and evaluates the cold chain logistics of rural agricultural products in Changsha County by using SWOT analysis method,and designs the cold chain logistics distribution scheme of rural agricultural products in Changsha County.(2)According to the cold chain logistics resource data of Changsha County crawled by python,based on the K-means + + clustering analysis algorithm of improving the initial center selection based on machine learning,the K value is determined by using the elbow method,and the K value with the highest score(k = 4)is selected by using the contour coefficient and karinsky halabas index method as the final number of township level agricultural products cold chain logistics distribution areas;And compared with the bottom-up Agnes algorithm,the K-means + +clustering analysis algorithm is finally selected to divide the distribution area,and the location of the distribution center in the distribution area is obtained.(3)According to the current situation and distribution characteristics of cold chain logistics of rural agricultural products in the county,three distribution modes are obtained: franchise store mode,concurrent store mode and cold chain self delivery cabinet mode;Aiming at the three distribution mode data of cold chain logistics in Changsha County crawled by python,based on three machine learning classification algorithms of Ada Boost,random forest and neural network,the R language is used to train and test the distribution mode data of cold chain logistics.The results show that the accuracy of Ada Boost algorithm is 87.5%,that of random forest is 79.17%,and that of neural network is 75%.(4)According to the Ada Boost algorithm with the best accuracy effect,the cold chain logistics distribution mode of agricultural products at the end of the village is selected.It is obtained that the cold chain self delivery mode is No.1 distribution area(longitude 113.14406206,latitude 28.11630481),No.3 distribution area(longitude113.29987512,latitude 28.14119025)and No.4 distribution area(longitude113.23661475,latitude 28.2671385),Distribution area 2(longitude 113.10221067,latitude 28.25818256)is a concurrent store mode.The results show that the classification results based on machine learning are consistent with the actual situation,which can provide some application value for the cold chain logistics distribution of rural agricultural products in Changsha County.