Multi-label classification models for sustainable flood retention basins

详细信息	查看全文 | 推荐本文 |

• 作者：Qinli Yang^a ; ^{Q.Yang-7@sms.ed.ac.uk} ; Junming Shao^b ; ^{shao@dbs.ifi.lmu.de} ; Miklas Scholz^a ; ^c ; ^{m.scholz@salford.ac.uk} ; ^{m.scholz@ed.ac.uk} ; Christian Boehm^b ; ^{boehm@dbs.ifi.lmu.de} ; Claudia Plant^d ; ^{cplant@fsu.edu}
• 关键词：Flood control ; Multi-label support vector machine ; Multi-label K-nearest neighbor ; Back-propagation for multi-label learning ; Multiple function ; Classification framework ; Water resources ; Landscape management ; Scotland ; Baden
• 刊名：Environmental Modelling and Software
• 出版年：2012
• 期刊代码：27_13648152
• 类别：env
• 出版时间：June, 2012
• 卷：32
• 期：Complete
• 页码：27-36
• 文件大小：1019 K

摘要

It is becoming good practice to prepare risk assessments of river basins and coastal areas on a global scale. The novel sustainable flood retention basin (SFRB) concept provides a rapid classification technique for impoundments, which have a pre-defined or potential role in flood defense. However, most SFRB do often perform multiple functions simultaneously and thus are associated with multiple SFRB types. Nevertheless, previous SFRB classification systems assign each SFRB to a specific type relying on its main function. To handle the problem, this study aims to comprehensively assess the multiple functions of SFRB with the help of multi-label classification. The popular multi-label classifiers multi-label support vector machine (MLSVM), multi-label K-nearest neighbor (MLKNN) and back-propagation for multi-label learning (BP-MLL) were applied to predict the types of SFRB based on two data sets (one from Scotland and one from Baden). Findings indicate that multi-label classification schemes provide deeper insights into all potential functions of SFRB and help planners and engineers to make better use of them.