基于XGBoost的机电安装工程工期预测模型研究

XGBoost-Based Duration Prediction Model for Mechanical and Electrical Installation Project

摘要: 随着机电安装工程复杂性和管理要求的不断提升，精准的工期预测成为优化项目管理和降低成本的关键。本文基于机电安装工程的实际数据，采用XGBoost算法构建工期预测模型，通过特征工程、模型训练和性能验证，实现了高精度的工期预测。研究结合数据样本特征，分析了施工效率、工序工程量和关键路径标记等关键因素对工期的影响，并通过与随机森林和线性回归模型的对比验证了XGBoost的优越性能。结果表明，该模型在均方误差(RMSE)和平均绝对误差(MAE)等指标上表现优异，特征重要性分析揭示了影响工期的核心因素，为工程项目优化提供了数据驱动的决策依据。

Abstract: With the increasing complexity and management requirements of Electrical Installation Project, accurate construction duration prediction has become a key factor in optimizing project management and reducing costs. This study develops a duration prediction model based on the XGBoost algorithm using real-world data from MEP installations. Through feature engineering, model training, and performance validation, the proposed model achieves high-accuracy predictions. By analyzing sample characteristics, the study identifies the impact of key factors such as construction efficiency, task workload, and critical path indicators on project duration. Comparative experiments with Random Forest and Linear Regression models demonstrate the superior performance of XGBoost. The results show that the model performs well in terms of Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). Additionally, feature importance analysis reveals the core factors affecting construction duration, providing data-driven insights for project optimization and decision-making.