Research On Customer's Classification And Recognition Of E-commerce Enterprise

With the rise of artificial intelligence technology,the retail business industry be in a new change based on data mining,"new retail" model in the traditional E-commerce industry quietly rise and spread,become the future development trend of E-commerce industry.The speed of information transmission,the increasingly demanding needs of consumers,increasingly saturated online shopping market,the traditional business development of the E-commerce industry are formed a challenge.Therefore,E-commerce industry enterprises to identify potential customers and the maintenance of existing users become business the key of concerns.Effective and accurate grasp of market preferences,the target market for accurate marketing is the direction of business efforts.In this paper,the customer recognition model is based on this purpose,for enterprises to clearly determine the target group and reduce the cost of business to find users.First,the user information is tagged according to the user portrait technology,and each user instance is marked with a specific label,and the tag group is divided according to the same tag.The label abstract extract feature attributes for the customer identification classification paving the way.By deducing the independence hypothesis of the naive Bayesian classification algorithm,the GINI coefficient is introduced to weight the feature attribute,and the influence degree of the attribute in the classification process is controlled according to the importance degree of the attribute feature.The improved Bayesian classification algorithm is verified by the UCI data set.The experimental results show that the Bayesian classification algorithm based on GINI coefficient is improved compared with the original Naive Bayesian algorithm.At the same time,in the empirical study,the accuracy calculation of the classification of the unbalanced data is modified,and the penalty variable is introduced based on the cost-sensitive learning idea,and the punishment of the sparse category is increased.The accuracy of the revised calculation is more focused on the role of the rare category on the overall data set,better reflect the true effect of the classification algorithm.