Application of Partial Least-square Regression to Engine Maintenance Project Cost Forecasting

摘要

The traditional multivariate regression methods cannot get satisfactory results as the impact of engine maintenance project cost forecasting by multiple and complex factors, and the sample data scarcity. This article proposes a Partial Leastsquare Regression model (PLSR, for short).PLSR has prominent advantages in dealing with small samples and variables with multiple relevance. The instance shows that PLSR is a better method, with lower error and higher accuracy. PLSR is a valid method for forecasting engine maintenance project cost and should be applied further.