Study on the Demand Forecasting of Hospital Stocks Based on Data Mining and BP Neural Networks

The demand forecasting of hospital stocks is a new management science with strong practicality, the basic principle of which is, on the foundation of keeping a high supply level of hospital clinical medicines, to apply mathematical and management methods to improve the accuracy of hospital stock demands forecasting, reduce the unreasonable stocks, accelerate cash flow, improve management measures of various aspects, and give full play to the hospital management functions in order to guarantee the safety and integrity of hospital medicines, and improve the hospital social and economic benefits. The paper firstly introduced the inventory control background of the Daping Hospital of Third Military Medical University (CQDP Hospital) and screened out the main influencing factors of the usage of 17GY scalp indwelling needles, and then applied data mining technologies including data cleaning, data integration and data transformation to pre-process the original data, obtaining the training samples for BP neural network, and finally built an inventory forecasting model based on BP neural networks, using the sample data for training to get the inventory forecasting network construction, and carried out the stock prediction in a certain period of the future, which provided basis for decision-making.