Safety Stock Prediction Model Based On The Fuzzy Decision Tree Algorithm

Safety stock is also known as the extra hold inventory, which is a very important link of supply chain in market. Identification of safety stock is directly influenced by several factors including the change of demand of inventory, the change of order interval, the change of delivery periodical delay, the shortage cost of inventory and the change of storage cost and so on. At the same time, safety stock shows a very sophisticated nonlinear feature. All of these factors make it extremely difficult to forecast the safety stock. So the study of safety stock has both importantly theoretical and significantly practical meanings.Aiming at the safety stock prediction problem, prediction model of safety stock based on fuzzy decision tree algorithm is proposed. The summary of the paper is as follows:(1) Analyze the key technologies of safety stock, mainly including the sketch of inventory and safety stock, identification of safety stock and the outline of determination method safety stock including utility value method, quantitative order method, ABC classification method, mathematical statistics method, marginal analysis, which also introduces the safety prediction principle, data mining technology;(2) it is to build safety stock prediction model based on fuzzy decision tree algorithm, to introduce the fuzzy set theory, to compare the decision tree to fuzzy decision tree and to collect, to manipulate and to analyze the test data;(3) Illustrate an example of the application of the model to predict the amount of safety stock based on fuzzy decision tree algorithm, which introduces the testing tools and environment, and realize the prediction model of safety stock standards based on the BP algorithm, prediction, and the amount of safety stock by using the traditional EOQ model. Finally, this paper will compare the safety stock prediction model based on the fuzzy decision tree introduced with the traditional EOQ safety stock prediction model and the prediction of safety inventory based on BP algorithm.The results show that: comparing the raw materials of Qkq using standard BP model and EOQ prediction model, the fuzzy decision tree model is close to the actual value, and the value of MAPE were less than the standard BP model and EOQ model 10.70 and 11.81 respectively,. To the finished goods WML using the raw materials Qkq employing standard BP model and the EOQ prediction model respectively, the result of the fuzzy decision tree model is close to the actual value, and the value of MAPE were less than the standard BP model and EOQ model 5.80 and 10.83 respectively, so it can conclude that prediction of the safety stock based on the fuzzy decision tree algorithm is high.