Rasearch On Order Of A Mall Based On Correlation Analysis

With the rapid development of e-commerce,the competition is more and more intense among the major e-commerce enterprises,the focus of competition is no longer limited to the price,enhance users' experience to being an important goal of the business platform.In the process of continuous improvement of people's quality of life,the price requirements will be reduced,but the requirements of the consumer experience is increasing.Formerly,people make decisions in the sales work mainly by virtue of experience and simple data analysis,but now,in the big data age,simple data analysis has been unable to meet the needs of businesses.If you could make full use of data mining technology,then you would mining customer characteristics,then making more efficient marketing strategy,and effectively enhance sales.In this thesis,we select all the orders data from June to July 16 of A Mall.The article mainly uses the method of correlation analysis to carry on the data mining to the order data,finding the correlation relationship,then contacting the situation to obtain the corresponding conclusion.Mainly,this thesis obtains three kinds of correlation relations,the first is the context correlation relation,which is the time,place,platform and the relationship between the goods can be advertised for the business to improve efficiency and make effective recommendations;The second relation is a commodity correlation relationship,which describe the relationship between goods and goods;The analysis can help businesses setting up activities,combining sales,setting prices and providing constructive advice for smart push.The last correlation relationship describes the loss of orders and order attributes are closely related.Then,according to the above three relations,the thesis puts forward the corresponding countermeasures and suggestions,and proposes the establishment of personalized intelligent push system for the first two kinds of related relations.Secondly,according to the order loss correlation,this thesis propose to establish the forecasting model of order loss,The model of order loss is established by random forest classification model,and the model is improved,evaluated and applied.