论文部分内容阅读
为了解决工程造价指数难以精确预测的难题,针对传统时间序列预测模型缺少内在固有信息而使得最终预测结果难以成功的缺点,从因变量的角度引入多变量时间序列的概念,并在ADF单位根检验和Johansen共整合基础上,证明多变量时间序列与工程造价指数存在协整关系,最后,结合支持向量机(SVM)预测算法,提出基于支持向量机(SVM)多变量时间序列回归预测算法,通过实验,结果表明,多变量时间序列可为造价指数预测提供更多的信息,预测算法的准确率较高且可行有效,具有实际利用价值,可为后续造价指数的预测和造价管理提供可靠的参考价值。
In order to solve the difficult problem that engineering cost index is difficult to predict accurately, aiming at the shortcomings that the traditional time series prediction model lacks inherent inherent information and makes the final prediction difficult to succeed, the concept of multivariable time series is introduced from the perspective of dependent variable. And Johansen, the cointegration relationship between multivariable time series and construction cost index is proved. Finally, combining with support vector machine (SVM) prediction algorithm, a multivariate time series regression prediction algorithm based on support vector machine (SVM) The experimental results show that multivariable time series can provide more information for the forecasting of cost index. The forecasting algorithm has high accuracy, feasible and effective, and has practical value, which can provide a reliable reference for the subsequent cost index forecasting and cost management. value.