<![CDATA[An inaccurate Owner’s Estimate Cost (OEC) often triggers procurement failures in the purchasing process, thereby affecting cost performance and the success of government capital expenditure projects in Indonesia from a construction management perspective. The OEC serves as the primary benchmark for assessing the reasonableness of bids in construction procurement; calculation errors may lead to ineffective cost control, financial mismanagement, and regulatory non-compliance. Therefore, this study aims to improve OEC accuracy by developing an Adaptive Neuro-Fuzzy Inference System (ANFIS) model supported by a Linear Regression (LR) algorithm to predict price fluctuations that influence cost planning and procurement decisions for public building projects. Project data from state-owned building construction and unit price analysis data for the 2021–2024 period were analyzed to predict 2025 price changes (addenda) for various construction work items. The proposed model achieved strong accuracy, with Root Mean Squared Error (RMSE) values of 0.0108–0.0333 and Mean Absolute Error (MAE) values of 0.0099–0.0261 across multiple work descriptions, indicating a good model fit. These findings confirm that the model outperforms comparable studies in terms of precision and interpretability, and can serve as a data-driven approach to strengthen cost management, estimate planning, and procurement decision-making in construction management.]]>
Copyrights © 2026
