Discovery And Recommendation For Telecommunication Package Using Intelligent Computing

With the competition between telecom operators becoming fierce, there are more and more telecom packages, which make customers dazzled. In many cases, customers don't know how to choose a suitable package for themselves. Personal experience is considered during design of telecom packages, and result in short life cycle and high management cost. Fortunately, intelligent age is coming. There are many practical problems are solved by intelligent algorithms. Then, more and more researchers are going to use intelligent algorithms in telecom package. However, there are not many researches about telecom intelligent algorithms and are usually confined to package concept, design rules, customer churn prediction and etc. Package optimization focuses on part of package attribute without considering the package impact on customers' choices. This research first proposes automatic optimization and recommendation of a telecom package. It balances the user's acceptance and operator's profits. The main contents are as following:This research first did data visualization analysis including picking up data from the original file and then checking the validity in data integrity and rationality to get the valid data. It analyzes user's consumption structure to find the weakness of their existing package. Based on the analysis results, it selects some attribute for setting up package model and customer model.This research uses particle swarm optimization to find a telecom package by comparing the advantages and disadvantages of genetic algorithm(GA) and particle swarm algorithm(PSO). In order to avoid local optimum, the parameters are adjusted dynamically. Operators gain the biggest profit by using the fitness value. With constant iteration, the inverse package is obtained. Considering the influence of all package attribute on customer's choice, this study calculates the proportion of user's original package and the inverse package. The comparison results are as a basis for finding potential customers.By using intelligent algorithm, this essay did mining analysis on customers data, operator inputs only several properties of the range of values, this system can optimize an inverse package automatically and select the candidate set for recommendation. In addition, it also adapts to an existing package for finding its candidate. With intelligent computing, it reduces blindness of artificial experience which will be a reference for operator's decision.Intelligent computing and recommendation for telecom package is an important means to promote intelligent service. It's good for operators to find out user's requirements and it also maintains relationships with the existing users while developing the new market. The promotion of a package will be more effective and accurate. The operator will improve their service and management level.