Research And Implemention Of Auto Parts Demand Prediction System Based On Kalman-ARMA Algorithm

With the continuous development of China's economy,the automobile industry has become closely related to people's life,and the sales of automobiles have increased.The increase of automobile sales also promotes the development of auto parts industry.Thus,the automobile after-sales service enterprises face more challenges.In order to be more competitive,after-sales service enterprises must be innovative and convenient,it is also essential to have high quality service and organizational management level.In order to improve the brand impact and customer service level,it is key to control the complete automobile spare parts.Thus,it is necessary to provide the scientific auto parts demand forecasting system.This thesis combines the whole automobile parts supply chain to study the actual demand forecasting system of auto parts.The prediction method adopted is ARMA,but due to the irregularity of the data,this thesis uses Kalman-ARMA combination prediction model to predict the demand.In this thesis,Kalman filter is used to process the data,then the ARMA iterative prediction model is established,and the ARMA parameters are adjusted by adaptive LMS algorithm.After the establishment of the algorithm model,the simulation is carried out and the performance of the simulation results is analyzed.Considering the correlation of auto service industry,the system needs to carry out data extraction from the external heterogeneous platform system,thus it is necessary to adopt the certain protocol technology to carry out the interaction of heterogeneous platform.This thesis uses WebService technology to carry out cross-platform data extraction,and combines Quartz to carry out automatic execution of data extraction tasks.In addition,this system uses the Druid database connection pool for performance monitoring to ensure the confidentiality of the data and the security of the system.Finally,every management module of the system are realized according to the business requirements,including statistical analysis management,data extraction management,demand prediction management,basic information management and system management.In addition,to verify the stability and practicability of the system,this thesis completes the overall system test,including functional test and loadrunner performance test.