學術產出-Periodical Articles

Article View/Open

Publication Export

Google ScholarTM

政大圖書館

Citation Infomation

題名 Cost-sensitive decision tree induction with label-dependent late constraints
作者 Kao, H.-P.;Tang, Kwei
唐揆
貢獻者 企管系
關鍵詞 Algorithms; Classification (of information); Costs; Decision trees; Diagnosis; Iterative methods; Classification tasks; Classification time; Constrained data mining; Conventional approach; Decision tree induction; Near-optimal solutions; Numerical experiments; Solution algorithms; Data mining
日期 2014
上傳時間 2-Jun-2015 17:11:07 (UTC+8)
摘要 Completion time requirements are often imposed on a classification task. In practice, the desired completion time for classifying a subject may depend on its label (target) value. For example, a timely diagnosis is important for an illness that requires immediate medical attention. It is common in medical diagnoses, therefore, to set completion times based on the severity level of the illness. In this study, we use "label-dependent" completion time requirements to formulate a new classification problem for cost-sensitive decision tree induction by adding "late constraints" to control the rate of tardy classifications for each label value. Adding the late constraints generalizes and enriches the decision tree induction problem, but also poses a challenge to developing an efficient solution algorithm because the conventional approach based on the "divide-and-conquer" strategy cannot be used. We develop a novel algorithm that relaxes the late constraints and iteratively solves a series of cost-sensitive decision tree problems under systematically-generated late penalties. The results of an extensive numerical experiment show that the proposed algorithm is effective in finding the optimal or a near-optimal solution. © 2014 INFORMS.
關聯 INFORMS Journal on Computing, 26(2), 238-252
資料類型 article
DOI http://dx.doi.org/10.1287/ijoc.2013.0560
dc.contributor 企管系
dc.creator (作者) Kao, H.-P.;Tang, Kwei
dc.creator (作者) 唐揆zh_TW
dc.date (日期) 2014
dc.date.accessioned 2-Jun-2015 17:11:07 (UTC+8)-
dc.date.available 2-Jun-2015 17:11:07 (UTC+8)-
dc.date.issued (上傳時間) 2-Jun-2015 17:11:07 (UTC+8)-
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/75527-
dc.description.abstract (摘要) Completion time requirements are often imposed on a classification task. In practice, the desired completion time for classifying a subject may depend on its label (target) value. For example, a timely diagnosis is important for an illness that requires immediate medical attention. It is common in medical diagnoses, therefore, to set completion times based on the severity level of the illness. In this study, we use "label-dependent" completion time requirements to formulate a new classification problem for cost-sensitive decision tree induction by adding "late constraints" to control the rate of tardy classifications for each label value. Adding the late constraints generalizes and enriches the decision tree induction problem, but also poses a challenge to developing an efficient solution algorithm because the conventional approach based on the "divide-and-conquer" strategy cannot be used. We develop a novel algorithm that relaxes the late constraints and iteratively solves a series of cost-sensitive decision tree problems under systematically-generated late penalties. The results of an extensive numerical experiment show that the proposed algorithm is effective in finding the optimal or a near-optimal solution. © 2014 INFORMS.
dc.format.extent 176 bytes-
dc.format.mimetype text/html-
dc.relation (關聯) INFORMS Journal on Computing, 26(2), 238-252
dc.subject (關鍵詞) Algorithms; Classification (of information); Costs; Decision trees; Diagnosis; Iterative methods; Classification tasks; Classification time; Constrained data mining; Conventional approach; Decision tree induction; Near-optimal solutions; Numerical experiments; Solution algorithms; Data mining
dc.title (題名) Cost-sensitive decision tree induction with label-dependent late constraints
dc.type (資料類型) articleen
dc.identifier.doi (DOI) 10.1287/ijoc.2013.0560
dc.doi.uri (DOI) http://dx.doi.org/10.1287/ijoc.2013.0560