Research And Design Of Shopping Basket Reorganization And Evolution And Visualization

Association rule mining is an unsupervised learning method.It mainly studies the knowledge model of customer buying commodity rules in transaction data.Association rules mining is the key technology of shopping basket analysis,especially in recent years,network retail by leaps and bounds,resulting in massive transaction data,shopping basket analysis put forward higher requirements.Shopping basket analysis for the supermarket stores and online mall to provide a variety of promotional reference to help group to inventory,optimize the layout of goods,increase turnover.Shopping basket analysis provides decision makers with fast,accurate,time-saving,and diverse information support.In the era of consumer upgrades,shopping basket analysis for business owners to improve the overall efficiency of the retail industry,to enhance international competitiveness,building a conservation-oriented society has important macro significance.In the shopping basket research process,through years of deep plowing retail business owners to communicate,get a conclusion that the traditional shopping basket analysis is simply a simple application of association rules,did not do a good combination of the actual situation,the results do not meet the application requirements.With the new retail industry booming,online electricity and offline physical store combination,to the shopping basket analysis brings great challenges.First of all,in the existing shopping basket analysis,the knot is a combination of some common goods.Although these shopping baskets are highly supported,they do not have to be analyzed to get the conclusion that the business is not much of a role.Second,the traditional shopping basket analysis of a great limitation is that it can not predict,usually only historical data for shopping basket analysis,not the shopping basket by time sequence to make evolution and prediction.Finally,the traditional shopping basket analysis only gives relatively simple results,did not realize the visualization of the results of the shopping basket analysis.In view of the above challenges,this paper proposes three innovative methods:1.Through the shopping basket cluster reorganization to filter more valuable shopping basket.The key to the reorganization of the shopping basket is to reorganize all similar shopping cartings with a clustering algorithm into a new,representative,quality shopping basket.Based on the core idea of division,this paper proposes a combination of hierarchical clustering and similarity methods to achieve the purpose of shopping basket reorganization.2.Shopping basket evolution is by digging analysis of historical transaction data to predict changes in shopping baskets.In this paper,the evolution of the parameters of the value and model research,and real data to simulate and detect.Taking into account the impact of time,promotion,holidays,and store surroundings on shopping baskets.3.Develop a shopping basket visualization interactive system that allows shopping basket analysis results to achieve friendly human-computer interaction and visualization.Validate the above method through real data,and provide a convenient analysis tool for retailers.