The Method And Application Of Medicine Sales Forecast For BY Company

With the continuous development of medicine circulation industry,continuous improvement of enterprises in the supply chain each link,medicine sales and the types of the deal increase rapidly.The traditional and single forecast method does not apply to medicine sales characteristic of the medicine circulation industry,uncertainty and running time too long of sales forecast method results is the main problem need to solve.Establishing accurate and rapid model to forecast medicine sales has practical significance for medicine circulation enterprise.BY company's many varieties and large-scale medicine sales forecast will be main research contents,based on the common sales forecast methods,in allusion to characteristics of medicine sales,the main problems existing in the course of inventory and sales for BY company,put forward a sales forecast model HCEA(Hierarchical Clustering-ELM-AdaBoost)about many kinds of medicine sales to forecast BY company's medicine sales.Then using BY company real 2476 medicine sales data to verify the validity of the model.The main contents of the paper can be summarized as the following:First of all,the paper will introduce BY the company's the status quo of sales and inventory management,analyze the variety of medicine sales forecast characteristics,main problems and difficulties,then state the clustering thought and the advantage of using the integrated model.Secondly,it will state the thought and modeling steps of hierarchical clustering algorithm,extreme learning machine(ELM)algorithm and AdaBoost algorithm in detail.Aiming at the limitation of current method and the problems of inventory and sales management for BY company to put forward a kind of above three algorithms together and suitable for many kinds of medicine sales forecast integration model called HCEA model.Finally,the paper will use BY company 2476 medicine real sales data to test HCEA model.The test results will be compared with AdaBoost-ELM model and AdaBoost-BP model,through the result comparison and t-test to verify the validity of HCEA model.