Highway construction cost index forecasting: a hybrid VMD–LSTM–GRU method

The Highway Construction Cost Index (HCCI) is a crucial metric for monitoring price trends in the highway construction industry, where accurate forecasting is essential for effective budgeting and resource allocation. However, the inherent volatility and complexity of HCCI data present significant challenges to predictive accuracy. This study addresses these challenges by proposing a novel hybrid method that integrates variational mode decomposition (VMD) with long short-term memory (LSTM) and gated recurrent unit (GRU) networks to enhance forecasting performance. The VMD technique decomposes the HCCI time series into intrinsic mode functions (IMFs), representing various signal frequency components. The LSTM model is employed to predict smooth IMF components while the GRU model handles the more volatile IMF components, ensuring robust performance across different data characteristics. The proposed VMD–LSTM–GRU framework was applied to the Texas HCCI dataset, demonstrating superior forecasting accuracy compared to conventional time series or deep learning approaches. The study advances the application of hybrid models in construction cost time series forecasting and introduces a new methodology for enhancing budget estimations and financial planning within the construction industry. By improving prediction accuracy, the VMD-LSTM-GRU framework offers significant potential for more reliable financial management and strategic planning in highway construction projects.