Selection of relevant features and examples in machine learning Export

@article{citeulike:402254, abstract = {In this survey, we review work in machine learning on methods for handling data sets containing large amounts of irrelevant information. We focus on two key issues: the problem of selecting relevant features, and the problem of selecting relevant examples. We describe the advances that have been made on these topics in both empirical and theoretical work in machine learning, and we present a general framework that we use to compare different methods. We close with some challenges for future work in this area.}, author = {Blum, A.}, citeulike-article-id = {402254}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=270626}, citeulike-linkout-1 = {http://dx.doi.org/10.1016/S0004-3702(97)00063-5}, citeulike-linkout-2 = {http://linkinghub.elsevier.com/retrieve/pii/S0004-3702(97)00063-5}, citeulike-linkout-3 = {http://www.ingentaconnect.com/content/els/00043702/1997/00000097/00000001/art00063}, citeulike-linkout-4 = {http://www.cs.tu.ac.th/classes/461/undergradcs202-/public/research/journal_version.pdf}, comment = {There is no guarantee that just because a feature is relevant , it will be necessarily useful to an algorithm (and vice versa) === Notation === * 

S is the sample set * D is the probability distribution of each element of S * c is the target function which maps an element from S to a class === Relevance to target 

Definition 1 === A feature x\_i is relevant to a target concept c if there exists a pair of examples A and B in the instance space such that A and B differ only in ther assignment to x\_i and c(A) != c(B). === Relevance as a complexity measure (used in Winnow)

 === Measures complexity of the target function. Practical goal is not to identify an irrelevant subset of features, but to perform well when the complexity is low.}, doi = {10.1016/S0004-3702(97)00063-5}, issn = {00043702}, journal = {Artificial Intelligence}, keywords = {feature-selection, relevance}, month = {December}, number = {1-2}, pages = {245--271}, posted-at = {2008-12-01 21:25:57}, priority = {0}, title = {Selection of relevant features and examples in machine learning}, url = {http://dx.doi.org/10.1016/S0004-3702(97)00063-5}, volume = {97}, year = {1997} }

TY - JOUR ID - citeulike:402254 L3 - citeulike-article-id:402254 N2 - In this survey, we review work in machine learning on methods for handling data sets containing large amounts of irrelevant information. We focus on two key issues: the problem of selecting relevant features, and the problem of selecting relevant examples. We describe the advances that have been made on these topics in both empirical and theoretical work in machine learning, and we present a general framework that we use to compare different methods. We close with some challenges for future work in this area. IS - 1-2 JF - Artificial Intelligence SN - 00043702 EP - 271 TI - Selection of relevant features and examples in machine learning VL - 97 SP - 245 KW - feature-selection KW - relevance N1 - There is no guarantee that just because a feature is relevant , it will be necessarily useful to an algorithm (and vice versa) === Notation === * 

S is the sample set * D is the probability distribution of each element of S * c is the target function which maps an element from S to a class === Relevance to target 

Definition 1 === A feature x_i is relevant to a target concept c if there exists a pair of examples A and B in the instance space such that A and B differ only in ther assignment to x_i and c(A) != c(B). === Relevance as a complexity measure (used in Winnow)

 === Measures complexity of the target function. Practical goal is not to identify an irrelevant subset of features, but to perform well when the complexity is low. AU - Blum, A PY - 1997/12// UR - http://dx.doi.org/10.1016/S0004-3702(97)00063-5 ER -