Research On The Compilation Of French Wine Hedonic Price Index Based On E-Commerce Data

With the rapid development of China’s economy,the wine industry,which is one of the traditional industries in China,has also developed rapidly.The growing Chinese wine market not only provides a broader space for the development of domestic wines,but also has made a huge attraction for foreign wine producers.They have invested their products in the Chinese market.The influx of imported wines has made the Chinese wine market competition increasingly fierce.If domestic wine producers and distributors want to gain a competitive advantage,the first task is to grasp the price changes accurately in the imported wine market.Therefore,it is necessary to prepare a single commodity price index for China’s imported wine.The development of e-commerce provides data support for the compilation of imported wine price index.However,due to the frequent transactions of e-commerce platforms and the rapid replacement of goods,old products will be eliminated and new products will appear every day.It is difficult to meet the "pure sample" conditions.At the same time,as a typical heterogeneous commodity,the quality of wine is constantly changing with time.It is difficult to maintain the homogeneity of the products when compiling the price index.In the SNA issued in 1993,the United Nations proposed an effective solution,which names hedonic price index compilation method,to solve this problem.The hedonic price model can effectively eliminate the influence of the quality change of heterogeneous products on the price,and compile the price index that reflects the "pure price change" of the commodity.Therefore,this paper summarizes the hedonic price method and applies it to the e-commerce French wine market.By selecting wine feature variables,constructing hedonic price model,estimating marginal hedonic price,and analyzing the important factors of wine price,this paper finally compiles the price index that reflects the "pure price change" of e-commerce French wines.Firstly,this paper collects the attribute information and sales data of wine products from March 2014 to February 2016 from various e-commerce platforms through web crawling technology.Since the quality of the data directly affects the accuracy of the price index compilation,this paper conducts detailed data pre-processing work on the e-commerce wine data,and finally obtained 30,470 effective transaction records,covering 7315 wine types,which are used for the compilation of e-commerce French wine hedonic price index.Secondly,based on the characteristics of the e-commerce wine database,this paper uses the random forest algorithm’s variable importance score evaluation to determine the seven characteristic factors.By dividing the above-mentioned characteristic factors,this text selected 18 characteristic variables to construct the hedonic price model.Finally,this text selects the logarithmic linear function form to construct the combined data hedonic price model for all sample data of 24 months,and the sectional data hedonic price model for each month’s sample data,and then compiles the Laplace sectional model fixed hedonic price index and the combined model fixed hedonic price index respectively.Moreover,by comparing two indices with the simple arithmetic average price index,the corresponding conclusions are drawn.Based on the above research,the text draws the following conclusions.First,the combined data model shows that grades,site categories,Bordeaux,Burgundy,Rhone Valley,Champagne,dried fruit aroma,oak aroma,spice aroma,year,Syrah and Pinot Noir have a positive impact on the price of e-commerce French wines.The size,Burgundy,Provence,plant aroma,floral and Cabernet Sauvignon have a negative impact on wine prices.Second,the sectional data model shows that consumers are concerned about the level of wine and whether the wines are carefully brewed in oak barrels.And compared to the integrated shopping website,consumers are more willing to pay higher prices for the wines of the liquor franchise website.Third,the index comparison shows that the combined data hedonic price index and the sectional data hedonic price index effectively eliminate the influence of the characteristics on the price of wine,and more truely reflect the trend of e-commerce French wine price.If we consider the sample size and the fluctuation of the rading structure of e-commerce platform,the combined data hedonic price index is more stable than the sectional data hedonic price index.The innovation of the article mainly includes the following two aspects:First,the research field of using the hedonic price theory for index compilation in China is relatively limited,mostly concentrated in real estate,electronic products and other fields.From a new perspective,this paper applies the hedonic price model to wine commodities for the first time,and compilates the wine hedonic price index accordingly,which complements and perfects the research of domestic hedonic price theory to some extent.Second,based on the study of wine hedonic price by foreign scholars and combined with the characteristics of e-commerce French wine transaction data,this paper uses the random forest algorithm’s variable importance score evaluation to determine the e-commerce French wine hedonic variable system,which provides a reference for domestic scholars to study the differences in wine commodity prices.The inadequacies of the article are mainly reflected in two aspects:First,the difference in the data published by each e-commerce platform makes the vacancy rate of the wine product attribute information higher,which leads to the incomplete selection of the characteristic factors affecting the wine price.The accuracy of the hedonic price model has had a certain impact.Second,since each country has its own wine hierarchy,it is difficult to directly compare different grades of wines in different countries.Considering that the grade is one of the important characteristic factors for the subsequent establishment of the hedonic price index model,this paper only compiled the hedonic price index for French wines,and failed to examine the price changes of the entire imported wine market in China.