查詢結果分析
來源資料
相關文獻
- Using Neural Networks to Estimate Replenishment Repair Parts Costs
- 以類神經網路估算建築工程成本之比較研究
- Neural Network Procedures for Taguchi's Dynamic Problems
- A Fast and Efficient Competitive Learning Design Algorithm Based on Weight Vector Training in Transform Domain
- 專家系統振動訊號圖型判別之研究
- 反傳遞模糊類神經網路於流量推估之應用
- 類神經網路(Neural Networks)的種類及其在影像處理上的應用
- C++Fuzzy類神經網路物件導向發展系統之建立
- 臺灣汽保費率之估計--對數線性費率模式與類神經網路之比較
- 運用類神經網路於股價指數之套利--以日經225指數為例
頁籤選單縮合
題名 | 以類神經網路估算建築工程成本之比較研究=Neural Network Application for Cost Estimating of Building Construction |
---|---|
作者 | 郭斯傑; 陳信夫; Guo, Sy-jye; Chen, Hsin-fu; |
期刊 | 建築學報 |
出版日期 | 19970900 |
卷期 | 22 1997.09[民86.09] |
頁次 | 頁81-94 |
分類號 | 441.3 |
語文 | chi |
關鍵詞 | 類神經網路; 成本估算; Neural network; Cost estimating; |
中文摘要 | 本研究以臺灣地區建築工程為估算的對象,蒐集某家甲級營造公司內部的成本資 料,以類神經網路進行各種狀況下的成本估算分析。本研究首先建構此種應用狀況下的最 佳類神經網路架構。選擇於建築工程概估階段,以經常考量的地質狀況、專案總坪數、地 上樓層數及地下樓層數為自變數,以估算出的每坪造價當作應變數,將九個實際案例分成不 同狀況,測試類神經網路的估算表現,並與傳統的工程分項統計估算法、專家估算、蒙地卡 羅模擬法、及多元線性迴歸估算法作成果比較。結果顯示,於八個訓練案例、一個測試案例 的情況下,類神經網路不論是在平均誤差平方和、誤差標準差或是誤差比率範圍,其表現均 優於其他各種估算模式。本研究證實了類神經網路於工程成本估算方面的準確性及適用性, 可提供更加準確的成本預估,減少估計的風險。 |
英文摘要 | In the competitive construction industry, cost estimating is a risk process that attempts to forecast the final cost of a future construction project. Recently researchers in artificial intelligence have provided a more potentially suitable tool, artificial neural network, for the task. Artificial neural network is reasonably suited for applications where prediction or forecasting is required based on incomplete,unrelated, or complex data. Thus, artificial neural networks have great potential for cost estimating. This paper develops an artificial neural network application on cost estimating of building construction to illustrate these potential techniques. The performances of the artificial neural networks are compared to other estimating methods:conventional statistic model, expert knowledge model, Monte Carlo simulation model and multilinear regression model with 9 completed building projects. The artificial neural network application predicts the final cost of these construction projects with smaller error of other estimating models. It outperforms the other methods and shows considerable promise in cost estimating for building construction. It suggests neural network of one hidden layer performs better than that has two or none layer. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。