Dynamic Forecasting of Hydroelectric Engineering Price Index Using Multidimensional Conditional Autoregressive Model: A Case Study in Southwest China

The price index of hydropower engineering plays a crucial role in cost management. Analyzing and predicting the trends of classified engineering price index using qualitative methods have limitations, as they may not be applicable to all price index. To address this issue, this study takes selected classified engineering price index in Southwest China as an example and establishes a multidimensional conditional autoregressive (CAR(n)) model based on the adjustment factor, i.e., price index. The results indicate the model demonstrates dynamic extrapolation properties with low model errors, particularly in short-term forecasting. Furthermore, the traditional approach of calculating price index through time-consuming weighted averages using adjustment factors is replaced with a new predictive method provided by this model. This model for calculating price index in specific projects or within a certain basin holds promising potential. Moreover, the determination of the forgetting factor in the model is related to many factors such as the duration of construction, price fluctuations, and economic development cycles. The selection of forgetting factor has a significant impact on the accuracy of the model, which is noted in modeling.