Estimating Distribution of Concrete Strength Using Quantile Regression Neural Network
學年 102
學期 2
發表日期 2014-07-16
作品名稱 Estimating Distribution of Concrete Strength Using Quantile Regression Neural Network
作品名稱（其他語言）
著者 I-Cheng Yeh
作品所屬單位
出版者
會議名稱 2014 4th International Conference on Civil Engineering, Architecture and Building Materials (CEABM 2014)
會議地點 中國, 海南島, 海口市
摘要 This paper is aimed at demonstrating the possibilities of adapting Quantile Regression Neural Network (QRNN) to estimate the distribution of compressive strength of high performance concrete (HPC). The database containing 1030 compressive strength data were used to evaluate QRNN. Each data includes the amounts of cement, blast furnace slag, fly ash, water, superplasticizer, coarse aggregate, fine aggregate (in kilograms per cubic meter), the age, and the compressive strength. This study led to the following conclusions: (1) The Quantile Regression Neural Networks can build accurate quantile models and estimate the distribution of compressive strength of HPC. (2) The various distributions of prediction of compressive strength of HPC show that the variance of the error is inconstant across observations, which imply that the prediction is heteroscedastic. (3) The logarithmic normal distribution may be more appropriate than normal distribution to fit the distribution of compressive strength of HPC. Since engineers should not assume that the variance of the error of prediction of compressive strength is constant, the ability of estimating the distribution of compressive strength of HPC is an important advantage of QRNN.
關鍵字 concrete;compressive strength;distribution;quantile regression;neural network
語言 zh_TW
收錄於
會議性質 兩岸
校內研討會地點 無
研討會時間 20140716~20140717
通訊作者
國別 CHN
公開徵稿
出版型式
出處 Applied Mechanics and Materials 584-586, pp.1017-1025
相關連結

機構典藏連結 ( http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/109519 )