基于粗糙集机器学习的全生命周期造价估算方法研究
Based on Rough Set Machine Learning of WLC Estimation Method

作者: 景晨光 :中铁第四勘察设计院,武汉; 段晓晨 :石家庄铁道大学,石家庄;

关键词: 全生命周期造价粗糙集机器学习 Whole Life Costing Rough Set Machine Learning

摘要:

本文利用粗糙集理论在知识发现上的优越性,结合机器学习的原理,以实际工程量清单样本为例,研究了历史数据不确定性影响下全生命周期造价的估算问题。在结合具体实例的基础上,给出了粗糙集从建模、有效数据筛选到决策规则生成、最终得出全生命周期造价结果的完整估算过程。本文尝试在全生命周期造价估算中引人粗糙集机器学习理论,从大量实测工程数据中优选出最有影响的因素,在保持决策属性和条件属性之间的依赖关系不变化的前提下,根据其等价关系寻找工程知识库中的冗余关系,从而简化决策表,确保其分类能力,约简掉联系较弱的因素,最后以粗糙集决策规则学习的形式实现造价预测。通过混淆矩阵交叉验证表明,应用粗糙集理论解决数据不确定性影响下的全生命周期造价估算是可行的。

Abstract: In this paper, rough set theory in knowledge discovery on the superiority of the combination of machine learning theory to the actual sample quantities, for example, the uncertainty of the historical data under the influence of life cycle cost estimation problem. In the light of the specific examples based on rough sets is given from the modeling, the effective data screening to decision rules generation, life cycle cost of the final results obtained the complete estimation process. This paper attempts to estimate life cycle cost of the introduction of rough set theory of machine learning, data from a large number of experimental works of the most influential factors in selection, the decision attribute and condition of maintaining the dependencies between attributes does not change the premise, according to engineering knowledge base to find the equivalence relations between the redundancy to simplify the decision table, to ensure that their classification ability, reduction factor out the weak links, and finally to study rough set decision rules are implemented cost forecast. Confusion matrix by cross-validation showed that the application of rough set theory under the influence of data uncertainty to resolve the full life cycle cost estimate is feasible.

文章引用: 景晨光 , 段晓晨 (2013) 基于粗糙集机器学习的全生命周期造价估算方法研究。 软件工程与应用, 2, 47-54. doi: 10.12677/SEA.2013.22009

参考文献

[1] 段晓晨, 张晋武, 李利军, 张健龙. 政府投资项目全面投资控制理论和方法研究[M]. 北京: 科学出版社, 2007: 12-30.

[2] 徐岳, 武同乐. 桥梁加固工程生命周期成本横向对比分析[J]. 长安大学学报(自然科学版), 2004, 24(3): 30-34.

[3] S. Yousefi, T. Hegazy, R. A. Capuruco, et al. System of multiple ANNs for online planning of numerous building improvements. Neurocomputing, 2008, 3(4): 346.

[4] 张勇. 粗糙集–神经网络智能系统在悬浮过程中的应用研究[D]. 大连: 大连理工大学, 2005.

[5] 孙士宝. 变精度粗糙集模型及其应用研究[D]. 四川: 西南交通大学, 2007.

[6] 交通部公路工程定额站. 公路工程工程量清单计量规则[M]. 长沙: 湖南省交通厅, 2005.

[7] 张云涛, 龚玲. 数据挖掘原理与技术[M]. 北京: 电子工业出版社, 2004.

[8] 程玉胜. Rosetta实验系统在机器学习中的应用[J]. 安庆师范学院学报(自然科学版), 2005, 2: 69-72.

分享
Top