Research On C Company Service Renewal Rate Forecast Modeling Based On Machine Learning

To become a century-old enterprise is the vision of C company,then a key indicator of becoming a century-old enterprise in the IT industry is that the proportion of recurring revenue accounts for more than 80% of total revenue.At present,C company's recurring revenue only accounts for less than 40%,and there is still a significant gap between reality and goal.The first thing that comes to recurring revenue is the service renewal income.So how to improve the service renewal rate of C company and bring annuity income to C company has become a critical strategic challenge for C company.My work focuses on applying data cleansing,data pre-processing,feature engineering and data modeling on the sales contract and service renewal contract big data in some previous financial years,obtained a service renewal machine learning model through training the data with XGBoost machine learning algorithm.The feature importance and influence analysis has been done based on the model,leading to some general advices on improving service renewal rate.Afterwards,based on predicting the latest sales contract data against the learned XGBoost model,adding a data analysis and visualization module,a service renewal rate prediction platform is built,which can give the top-down list of the sales contracts based on the probability of losing the service renewal,and detailed root cause analysis and action recommendations can be given to the specific contract.Through understanding the probability of contract renewal and the renewal drivers,thereby improving the service sales decision efficiency,optimizing the service coverage,increasing service renewal rate of C company and bringing lasting incremental income to C company.