Research On Promotion Selection Of E-commerce Platform Based On Data Mining

With the rapid development of China's Internet e-commerce,the phenomenon of nationwide online shopping already exists,but with the development of China's e-commerce system,user dividends begin to disappear,and the launch of promotional festivals and other forms become the key means for e-commerce platforms to reduce customer acquisition costs and increase revenue,and in this process,how to reasonably choose products for promotion is a very important link for the platform.However,in the daily promotion selection of e-commerce platform,due to the uneven operation level of operators and the large number of commodities they are responsible for,when the platform promotes commodities,it is unable to select the best among the best,or select the best locally,and then it is unable to make a comprehensive layout,resulting in the unreasonable allocation of e-commerce platform resources,which makes the promotion effect of e-commerce platform poor.Therefore,it is an urgent problem for e-commerce platform to explore how to complete the one-stop online selection strategy when doing promotional activities,and further analyze the factors affecting its promotion.This thesis discusses the current situation of e-commerce platform market and promotion in China,analyzes the factors that may affect the selection of e-commerce platform products in promotion,such as one week's preferential intensity and the characteristics of users who choose to buy Promotional products,and then discusses the construction of intelligent selection model of e-commerce platform using big data technology.Then,this thesis applies the transaction history data of a large-scale well-known domestic e-commerce platform and the commodity and activity information data of the operation end,and constructs the promotion selection scheme for realizing the intelligent operation of e-commerce platform by means of data mining and machine learning algorithm.Including the prediction of Gmv of the total transaction amount of platform commodities during the promotion period by using random forest and xgboost algorithm.From the GMV prediction results,top N products were selected as the main candidate products,and the word embedding tool word2 vec method in the field of natural language processing is used to identify the main candidate products corresponding to the related products,and than the related product list was obtained for personnels make selections in promotional scenarios.Based on the product data and price data,and the use of big data technology,the online intelligent selection for promotion scenarios in the e-commerce platform is realized,making event planning and product selection more efficient while increasing the return on investment of sales.