LDR | | 00000nam u2200205 4500 |
001 | | 000000432729 |
005 | | 20200224134438 |
008 | | 200131s2019 ||||||||||||||||| ||eng d |
020 | |
▼a 9781687983152 |
035 | |
▼a (MiAaPQ)AAI13884183 |
040 | |
▼a MiAaPQ
▼c MiAaPQ
▼d 247004 |
082 | 0 |
▼a 310 |
100 | 1 |
▼a Wang, Long. |
245 | 10 |
▼a Cost-sensitive Boosted ROC Classification Trees. |
260 | |
▼a [S.l.]:
▼b State University of New York at Stony Brook.,
▼c 2019. |
260 | 1 |
▼a Ann Arbor:
▼b ProQuest Dissertations & Theses,
▼c 2019. |
300 | |
▼a 108 p. |
500 | |
▼a Source: Dissertations Abstracts International, Volume: 81-05, Section: B. |
500 | |
▼a Advisor: Zhu, Wei. |
502 | 1 |
▼a Thesis (Ph.D.)--State University of New York at Stony Brook, 2019. |
506 | |
▼a This item must not be sold to any third party vendors. |
520 | |
▼a Classification is one of the most important problems in machine learning. Plenty of successful algorithms have been developed for classification of balanced data. However, challenges arise when we have severely imbalanced data among different classes, that is commonplace in real-world applications. Utilizing the receiver operating characteristic curve (ROC), our group had recently developed the ROC classification tree to effectively identify minority class observations by using the area under the ROC curve and the harmonic mean of sensitivity and specificity to determine the node splitting criteria. While successful at classifying the minority classes, the ROC classification tree can be further improved in terms of its overall classification performance. On one hand, ensemble methods such as boosting and random forest would combine individual classifiers such as classification trees to improve classification performance. On the other hand, cost-sensitive learning techniques have also been successful in solving the class imbalance problem by taking the misclassification costs of different classes into considerations. In this dissertation, we propose the novel cost-sensitive boosted ROC classification trees incorporating the ROC classification tree as base classifier and the cost-sensitive boosting algorithm as the ensemble method. The proposed method is applied to both bi-class and multi-class imbalance classification problems. The classification performance of this new classifier is evaluated using several datasets from the UCI machine learning repository and comparisons are made to other classifiers including the ROC classification tree, and the random forest. |
590 | |
▼a School code: 0771. |
650 | 4 |
▼a Statistics. |
690 | |
▼a 0463 |
710 | 20 |
▼a State University of New York at Stony Brook.
▼b Applied Mathematics and Statistics. |
773 | 0 |
▼t Dissertations Abstracts International
▼g 81-05B. |
773 | |
▼t Dissertation Abstract International |
790 | |
▼a 0771 |
791 | |
▼a Ph.D. |
792 | |
▼a 2019 |
793 | |
▼a English |
856 | 40 |
▼u http://www.riss.kr/pdu/ddodLink.do?id=T15491351
▼n KERIS
▼z 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |
980 | |
▼a 202002
▼f 2020 |
990 | |
▼a ***1008102 |
991 | |
▼a E-BOOK |