A Study On The Management Of Retail Business Forecasting For Sinopec M Branch

Along with the gradual opening of Chinese refined oil products market after entering WTO, the market participants increased. Oil product sales company need to survive and develop in the environment of competition, and the market competition is becoming more direct and more serious. Sinopec is also suffering an unprecedented impact and challenges, whose market share is gradually lost. Product oil retail management is a complicated system. Retail sales forecasting analysis and research is the front-end work of the retail management. Accurate prediction of product oil retail enterprises can give guidance to strategic planning and the development of retail plans and optimize the allocation of resources etc.This paper selects Sinopec Sales Co. Ltd. M branch as an example. Firstly, it introduces the basic situation and the general retail management of M branch, and analyzes the current situation of Retail Forecast Management deeply, then finds out the problems exist in retail forecast process and method, and discusses the importance of scientific and accurate prediction. Secondly, through a comprehensive and systematic analysis of the factors affecting the retail sales of refined oil products, including the macro environmental factors, the industry environmental factors and the internal factors etc, using both qualitative analysis and quantitative analysis, to understand the refined oil retail marker further, has formed a set of scientific and complete influence factors system for the modeling data. Thirdly, the retail forecast model is established based on an analysis on the factors. From the forecasting purpose, this paper tries to use two kinds of data mining methods, linear regression and neural network to model. The two methods have their own advantages, to find out the intrinsic link between the refined oil retail sales and the influencing factors, and then explain the model, evaluation and application. Because of the remaining 5 indexes of linear regression backward regression is included in the 6 most important indicators of former neural network, the results of two models are basically the same. Finally, on the basis of database modeling and according to the current work, it puts forward the suggestions and countermeasures of oil products retail forecast management, that through the arrangement of the monitoring points and the establishment of large data analysis, to obtain a scientific and accurate forecasting method.