摘要
为提高建筑工程造价预测的准确性和效率,研究提出了一种工程造价预测模型—改进粒子群优化的双向长短期记忆网络。研究首先通过利用灰色关联分析法对特征指标进行筛选,然后引入了混合注意力机制和阶段性调节策略,以提升模型的预测准确性。实验结果显示,与现有先进模型相比,研究提出的算法在预测精度上具有明显优势,其绝对误差区间为-18~17元。上述结果表明,研究提出的建筑工程造价预测模型能实现对建筑造价的准确预测,具有重要的实用价值。
To improve the accuracy and efficiency of construction project cost prediction,this paper proposes an engineering cost prediction model--improved particle swarm optimization bidirectional long and short term memory network.The study first screened the characteristic indicators using the grey relational analysis method,then introduced a hybrid attention mechanism and phased adjustment strategies to improve the model's predictive accuracy.Experimental results show that compared to existing advanced models,the proposed algorithm demonstrates a significant advantage in prediction accuracy,with an absolute error range of-18 to 17 yuan.These findings indicate that the proposed construction project cost forecasting model can accurately predict construction costs,making it highly practical.
作者
李正焜
胡旭冉
章凤
LI Zhengkun;HU Xuran;ZHANG Feng(Department of Grain Engineering,Anhui Grain Engineering Vocational College,Hefei 230000,China)
出处
《佳木斯大学学报(自然科学版)》
2025年第11期58-61,共4页
Journal of Jiamusi University:Natural Science Edition
基金
2023年度安徽省质量工程项目(2023cxtd243)。
