Research On Commodity Selection Of Online Group Buying Based On Data Mining

Online group buying Refers to the latest online shopping model that a large group of people buying the same item on the same website in a limited period of time to get discount prices from retailer. Online group buying is a hot topic and the market develops quickly. Nowadays group buying websites have to intervene a lot of manpower for product review when facing the large number of application. It also may excessively dependent on the former experience. This may also lead to poor User experience and website user loss due to commodity quality and service problem. How to choose the suitable merchant and commodity is becoming the key of the future of group buying websites and the whole industry.This paper presents a method to evaluate online group buying sales level based on data mining. Classify the sales level by using TSC algorithms. Analyze the impact of sales level and choose the suitable input variable by the method of ANOVA Analysis, correlation analysis, likelihood-ratio tests and so on. Lastly, using the cost matrix, pruning confidence adjustment and modified boosting method to build the decision tree model on the basis of C4.5.After modeling, also evaluate the rate of accuracy and interpret the rule we generated. The method of this paper can Extended to other category to support the decision for commodity selection. Also able to improve the automate selection as well as the data management of websites.In this paper, we analysis the data under the clothing, bags and shoes category of a large group buying website in2012second quarter based on data mining method. We found the important factors affecting sales levels:price, discount, in barn or not and so on. We also built the decision tree classification rule.