Bacterial Heuristic Multi-objective Feature Selection Method For Customer Segmentation

With the rapid development of e-commerce and the intensification of competition in the global market economy,traditional enterprises' competition model which is based on 4Ps marketing theory(product,price,place and promotion)has been gradually replaced by the business philosophy based on customer relations.In the long-term business process,the company has accumulated a large amount of customer data,which can truly and objectively reflect the customer's consumption pattern and potential demand.Effectively combining intelligent customer segmentation technology with the company's rich customer data resources can help companies prioritize customers and develop faster,more accurate and effective business management strategies.Under the auspices of the National Natural Science Foundation of China(71571120),this paper firstly reviews and summarizes customer segmentation problems and feature selection methods.Then,it constructs a customer feature selection optimization model with customer classification error rate minimization and feature number minimization as two objectives.At the same time,a new type of multi-objective bacterial heuristic is designed.Finally,the developed optimization technique is applied to multi-objective customer segmentation problem instances.This paper carries out the following three aspects of research:Firstly,a multi-objective customer feature selection optimization model is established.With the objectives of minimizing customer classification error rate and minimizing the number of features,this paper aims to construct a multi-objective customer feature selection optimization model to select the most suitable features for analyzing customer consumption behavior,so as to provide guidance for the development and implementation of customer relationship management strategy.Secondly,a new multi-objective bacterial heuristic algorithm is designed for customer segmentation problems.As a new type of swarm intelligence optimization method,bacterial heuristic algorithm has been successfully applied to multiple optimization fields due to its powerful parallel computing ability and strong robustness.This paper extends the original single-objective bacterial heuristic algorithm to multiobjective algorithm.This paper also introduces the bacterial health evaluation mechanism to ensure the quality of the optimized solution;improves the external archive storage mechanism to achieve efficient storage of the solution obtained in the optimization process;designs a "hidden order"-oriented information exchange mechanism to improve the efficiency and convergence rate of the algorithm;introduces the boundary control mechanism to ensure the feasibility and diversity of the solution.In order to test the effectiveness of the improved algorithm,this paper introduces four benchmark data sets,and compares the proposed algorithm with the other five representative multi-objective methods.Thirdly,develop optimization techniques for customer segmentation model solving.For multi-objective customer segmentation problem,this paper aims to solve the customer segmentation optimization model by designing the bacterial heuristic multi-objective feature selection method.Each bacterium represents a potential feature combination scheme,which is constantly optimized in an iterative search.This paper selects the Australian Customer Approval Data set,the German Credit Data set,and Default of Credit Card Clients Data set as the experimental data sets of the customer segmentation,and carries out a comparative study with other multi-objective heuristic feature selection optimization methods on this issue.The experimental results verify the effectiveness of the proposed optimization method in solving multi-objective customer segmentation problems.