Research on Project Cost Prediction Model of University Construction Projects Based on Audit Perspective
郭艳利 GUO Yan-li
(北京邮电大学审计处,北京 100876)
(Audit Office of Beijing University of Posts and Telecommunications,Beijing 100876,China)
摘要:为加强工程审计智能化建设,高效利用已有工程审计数据估算工程造价,对于新时期背景下完成工程审计转型具有重要意义。本文以高校已建成学生宿舍、食堂、教学楼等为研究对象,系统分析了影响建筑工程造价的主要因素,通过相关性分析提出了预测工程造价的特征指标,构建了工程造价 BP 神经网络预测模型,以预测新建项目工程造价。文章在真实案例训练集上训练模型,并不断优化调整模型参数,对实际已发生的工程造价测试集数据进行验证,结果表明工程造价预测误差率不超过 8.19%,为快速、精准预测工程造价提供了新的思路和方法。此外,通过对已竣工项目结算审计成果进行科学分析,有利于促进审计成果转化和应用,发挥审计宏观层面监督效能,为建设单位新建项目科学、高效决策提供数据支撑。
Abstract: To enhance the intelligent construction of engineering audits and efficiently utilize existing engineering audit data for cost estimation, it is crucial to achieve engineering audit transformation in the new era. This study focuses on completed projects in universities, such as student dormitories, canteens, and teaching buildings. We systematically analyze the main factors affecting construction costs and identify predictive indicators through correlation analysis. A BP neural network model is constructed to predict the construction costs of new projects. The model is trained on a real-case training set and continuously optimized by adjusting its parameters. Validation on actual project cost data shows a prediction error rate of no more than 8.19%, providing a new approach for rapid and accurate cost prediction. Additionally, scientific analysis of completed project audit results promotes the transformation and application of audit outcomes, enhancing the macro-level supervision effectiveness of audits. This supports construction units in making scientific and efficient decisions for new projects.
关键词:高校基建;工程审计;特征指标;BP 神经网络;工程造价预测
Key words: university infrastructure;engineering audit;feature indicators;BP neural network;construction cost prediction
中图分类号:TU723.3 文献标识码:A 文章编号:1006-4311(2024)28-036-03 doi:10.3969/j.issn.1006-4311.2024.28.011
|