The PMI, the T‐Bill and Inventories: A Comparative Analysis of Neural Network and Regression Forecasts
比较了回归和神经网络方法预测采购经理指数（PMI）的效果，发现短期利率领先PMI变化10个月，且PMI可预测8个月后的实际库存变化，对供应链管理者有参考价值。
SUMMARY The PMI is widely used as an indicator of economic trends, and as a short‐term forecaster of several important lagging output variables. The PMI's importance has led to several attempts at forecasting its direction. This research builds on this foundation by attempting to forecast the PMI through the use of regression and neural network methodology. Findings indicate that short‐term interest rates lead and forecast changes in the PMI by 10 months. Additionally, the paper focuses on the PMI as a predictor. Of the many variables the PMI can fit or predict, results reveal inventories as the most relevant example for supply chain managers, showing the PMI is a predictor of real inventory changes 8 months out.