Analysis And Research On Spatial Price Transmission Of Agricultural Product Based On Internet

The problem of difficult to buy and sell agricultural products is becoming more prominent in China,which has seriously affected the farmers' income and stability of agricultural economic order.As an indispensable component of the price mechanism,the price transmission of agricultural products is very important for the timely discovery of the phenomenon of difficult to buy and sell in the process of commodity trading.At present,the price transmission study of agricultural products is the main use of agricultural statistical yearbook data of countries and provinces,there is a serious lag and at a coarse-grained level,which fails to find the price transmission laws of agricultural products in different cities quickly.With the rapid development of information technology,the Internet has become the most important source of information.There are more than 30000 agricultural business platforms which update announcements,supply and demand,price and other news every day,data is in real time and price can be refined to each city farmers' market level.So it is likely to be more accurate for finding spatial price transmission path of agricultural products by using the platforms data.Based on the background of the Internet big data,direction and period of the price transmission of agricultural products in different regions were studied.Specifically,the following 3 parts were included.(1)Due to the lack of a unified description of Internet information,data is redundant and expressions are inconsistent.Firstly,the agricultural product name was recognized and classified on the semantic level automatically.Traditional information extraction method based on Conditional Random Field(CRF)relies on sample corpus,it is expensive to label corpus manually and the extraction precision is low without considering the semantic features in processing agricultural product name and categories recognition.In order to solve this problem,a method of agricultural product name and categories recognition based on CRF and agricultural ontology was proposed.By using word,part of speech,geographical attributes and agricultural ontology concept features,a total of 4 groups of comparative experiments were completed and 7 categories were identified.The experimental results show that the overall precision,recall and F-score of the open test are increased by 10.20%,59.78%and 35.17%respectively by adding ontology concepts,therefore the semantic level acquisition about the agricultural products information on the internet were realized effectively.(2)Take tomato as example,a total of 138 price time series of 26 cities including Beijing,Shanghai,Chongqing from January 11,2016 to February 27,2017 with 3 days as a cycle were selected.Take every city as the center and 300 kilometers as the radius,the cities.were divided into 8 regions.ADF unit root test for judging stationarity,Johansen cointegration test for judging long equilibrium and Granger causality test for judging transmission direction were finished on the regional tomato price data.Considering both spatial and temporal factors,the reasonable lag period was estimated and Granger path was obtained of spatial price transmission which are conductive to reducing wide fluctuation of price.(3)Based on the organized relevant data of agricultural products of spatial price transmission path of tomato,visual system was developed which can realize the visualization of different dimensions such as time,price or display area and improve the decision interaction.