| From the 70's until now,prices of agricultural products in China has been in a state of fluctuation,especially after 20 08,the period of the price fluctuations of agricultural products has been shortening,has a asymmetry of fluctuating.Rise rate is higher than drop rate,which is called the state of spike.But price fluctuations of agricultural products not only has a direct impact on the farmers' income and farmers'production enthusiasm,but also has an influence on the people's daily life and vital interests.The information asymmetry is considered as the key factor for price fluctuations of agricultural products.Especially,it is hard for farmer to get the information of real-time accurate price and sales.Now,the agricultural website is of varying quality,most of them are served not for the farmer but for the goverment,so,we must take an authoritative agricultural website as an object for the web information collection,then we can use the time series prediction method to predict the future trend of price,it will be a very good decision-making basis.The main content of this paper:1.This paper summarizes the relevant technology of the web crawler,such as open source tools,Heritrix.Relevant technology of network information extracting.The agricultural information collection is based on the national database of agricultural products,crawling language is php/python,matching the HTML source code of a web page by regular expressions.First of all,crawling agricultural species,which are grain and oil,vegetables,fruits,aquatic products and animal by-products.The second,crawling agricultural products name and its related information's urls.The last,analysizing website source code,writting the code to crawl the agricultural products market,agricultural prices,the release date of agricultural prices.After a crawl above,obtaining the completed price quotes of agricultural products,a total of more than 50 ten thousand data,and the data is increaseing everyday.2.Comparing Heritrix acquisition process with PHP/python code acquisition process,introducing their advantages of two methods.3.Regarding the carrots' price data in June 2006,2007,2008,2009,2010,2012,2013 in wholesale markets of agricultural products in fengtai district of Beijing xinfadi as the research object,predicting 2014 prices.Prediction methods are average prediction method,the weighted arithmetic mean method,the growth in average time series prediction method,market forecast method for the average development speed of time series,a moving average forecast method,quadratic moving average forecasting method.Then,through the error square to judge the advantages and disadvantages of various methods.4.The agricultural information collection and release platform is divided into agricultural information collection module,agricultural information query,comparison module,a module to request to purchase agricultural products,login,registration module and management module.Among them,price comparison chart and the price trendency chart are drawed by highcharts API.The interface of the agricultural information collection and release platform based on the B/S is completed by rewriting the BootStrap toolkit,which gives user a concise and beautiful feeling,that is a very good user experience.General search engines can't understand users' search purpose well,returning a large number of web content that users don't care about,web information collection is aimed at a particular site in this paper,compared to the traditional search engines,be more specific and more targeted,meet the needs of users well. |
PDF Full Text Request