Variable Selection Of Longitudinal Data Model And Its Application In Network Marketing

In the context of large Internet data,the longitudinal data because of its can be effectively combined cross section data and time series data,in the network marketing occupies a very important position.Especially in the network marketing is also often due to high dimensional data sparseness,resulting in there is a significant difference between the data processing methods in the high dimensional space and the low dimensional space.It is necessary to further solve and study.Traditional technology in the large data environment can not be a good study of high dimensional data.Therefore,this paper combines the new variable selection method with the traditional longitudinal data,the specific research contents and results are as follows:First of all,the Elastic Net method is applied to longitudinal data model which appears in network marketing.It not only makes us better understand the impact of big data on a variety of marketing activities,but also allows companies to better play its effectiveness.The Elastic Net estimation of longitudinal data model is established and proved that this model has the nature of group effect.The data simulation proves that the Elastic Net method can select all the strongly correlated variables into the model while the Lasso method can not.Secondly,although the Elastic Net method has some effect on dealing with strongly correlated variable group problems,it still does not have Oracle properties as the Lasso method.In this paper,the Adaptive Elastic Net method is combined with the longitudinal data model to prove that the Adaptive Elastic Net method can deal with the strongly related variables and zero variables,that is,the group effect nature and the Oracle nature.And the numerical simulation shows that it is superior to the Lasso method and the Elastic Net method.Finally,we take advantage of the actual example of ad click through rate in network marketing,combining the longitudinal data with the ad click through rate,using the Elastic Net method to variable selection,and selecting important keywords to better improve ad click through rate.It is shown that the Elastic Net method is feasible for the longitudinal data model,and the fitting effect and prediction ability of the model are stronger than the traditional longitudinal data model.At the same time it also implements the Elastic Net method in the network marketing application.