The Wrapper Approach
作者： Ron Kohavi , George H. John
DOI: 10.1007/978-1-4615-5725-8_3
关键词: Selection (relational algebra) 、 Focus (optics) 、 Relevance (information retrieval) 、 Artificial intelligence 、 Machine learning 、 Feature (computer vision) 、 Filter (mathematics) 、 Relation (database) 、 Feature selection 、 Decision tree 、 Computer science
摘要: In the feature subset selection problem, a learning algorithm is faced with problem of selecting relevant features upon which to focus its attention, while ignoring rest. To achieve best possible performance particular on training set, method should consider how and set interact. We explore relation between optimal relevance. The wrapper searches for an tailored domain. compare approach induction without Relief, filter selection. Improvement in accuracy achieved some datasets two families algorithms used: decision trees Naive-Bayes. addition, subsets selected by are significantly smaller than original used algorithms, thus producing more comprehensible models.
springer.com
core.ac.uk
sci-hub.st 参考文章(43)
Thomas G. Dietterich, Hussein Almuallim, Learning with many irrelevant features national conference on artificial intelligence. pp. 547- 552 ,(1991)
Ron Kohavi, James Kelly, Cliff Brunk, MineSet: an integrated system for data mining knowledge discovery and data mining. pp. 135- 138 ,(1997)
Ron Kohavi, Dan Sommerfield, Feature subset selection using the wrapper method: overfltting and dynamic search space topology knowledge discovery and data mining. pp. 192- 197 ,(1995)
Avrim L. Blum, Ronald L. Rivest, Original Contribution: Training a 3-node neural network is NP-complete Neural Networks. ,vol. 5, pp. 117- 127 ,(1992) , 10.1016/S0893-6080(05)80010-3
Matt Ginsberg, Essentials of Artificial Intelligence ,(1993)
Thomas G. Dietterich, Machine-Learning Research Ai Magazine. ,vol. 18, pp. 97- 136 ,(1997) , 10.1609/AIMAG.V18I4.1324
Cohn L. Mallows, More Comments onCp Technometrics. ,vol. 37, pp. 362- 372 ,(1995) , 10.1080/00401706.1995.10484370
Michael J. Pazzani, Pedro M. Domingos, Beyond Independence: Conditions for the Optimality of the Simple Bayesian Classifier. international conference on machine learning. pp. 105- 112 ,(1996)
Ron Kohavi, Dan Sommerfield, James Dougherty, Data Mining using MLC++, A Machine Learning Library in C++ international conference on tools with artificial intelligence. pp. 234- 245 ,(1996)
Ron Kohavi, David Wolpert, Bias plus variance decomposition for zero-one loss functions international conference on machine learning. pp. 275- 283 ,(1996)