KNOWLEDGE DISCOVERY AND DATA MINING AGENTS

30 Jonathan I. Maletic and Andrian MarcusBallou, D. P. & Tayi, G. K. Enhancing Data Quality in Data Warehouse Environments, Com-munications of the ACM 1999; 42(1):73-78.Barnett, V. & Lewis, T., Outliers in Statistical Data. John Wiley and Sons, 1994[r]

of the KDD/DM community in research and practice. This handbook evolved fromthese experiences.The first edition of the handbook, which was published five years ago, was ex-tremely well received by the data mining research and development communities.The field of data[r]

Bay S.D., Schwabacher M., ”Mining distance-based outliers in near linear time with ran-domization and a simple pruning rule,” In Proc. of the ninth ACM-SIGKDD Conferenceon Knowledge Discovery and Data Mining, Washington, DC, USA, 2003.Ben-Gal I., Mora[r]

to be only MCAR or MAR, and the set of Bayesian networks is limited to thosein which the partially observed variable is a child of the other variables. Researchis needed to extend these results to the more general graphical structures, in whichseveral variables can be partially observed an[r]

sensitive medical information retrieval, The 11th World Congress on Medical Informat-ics (MEDINFO 2004), San Francisco, CA, September 2004, IOS Press, pp. 282–286.Blum P. and Langley, P. Selection Of Relevant Features And Examples In Machine Learning,Artificial Intelligence, 1997;97: 24[r]

and den Bussche, 2000). Ng et al. have listed a large collection of constraints andclassified them into several classes for which different optimization techniques couldbe used (Ng et al., 1998). The most studied classes or the class of so-called anti-monotone constraints, as is the minimal su[r]

170 Lior Rokach and Oded Maimonsuch as: supervised learning lr6,lr12, lr15, unsupervised learning lr13,lr8,lr5,lr16 andgenetic algorithms lr17,lr11,lr1,lr4.ReferencesAlmuallim H., An Efficient Algorithm for Optimal Pruning of Decision Trees. ArtificialIntelligence 83(2): 347-362, 1996.Almuallim[r]

the training set, and the question of whether it can be improved, and if so how, isan open and important one. Part of the answer to this question is to determine theminimum error achievable by any classifier in the application domain (known as theoptimal Bayes error). If existing[r]

872 Slava Kisilevich, Florian Mansmann, Mirco Nanni, Salvatore RinzivilloBaglioni M, Antonio Fernandes de Macedo J, Renso C, Trasarti R, Wachowicz M (2009)Towards semantic interpretation of movement behavior. Advances in GIScience pp 271–288Berndt DJ, Clifford J (1996) Finding patterns in time serie[r]

130 Irad Ben-GalRunger G., Willemain T., ”Model-based and Model-free Control of Autocorrelated Pro-cesses,” Journal of Quality Technology, 27 (4), 283-292, 1995.Ruts I., Rousseeuw P., ”Computing Depth Contours of Bivariate Point Clouds,” In Compu-tational Statistics and Data Ana[r]

play an important role in the process of scientific discovery. A system may discoversalient features in the input data whose importance was not previously recognized. Ifthe representations formed by the inducer are comprehensible, then these discoveriescan be made accessible to human re[r]

of its leaves can be transformed into a rule simply by conjoining the tests along thepath to form the antecedent part, and taking the leaf’s class prediction as the class9 Classification Trees 151value. For example, one of the paths in Figure 9.1 can be transformed into the rule:“If customer a[r]

118 Irad Ben-Galathlete performance analysis, and other data-mining tasks (Hawkins, 1980, Barnettand Lewis, 1994, Ruts and Rousseeuw, 1996, Fawcett and Provost, 1997, Johnsonet al., 1998, Penny and Jolliffe, 2001,Acuna and Rodriguez, 2004, Lu et al.,[r]

the issue that different discretization strategies are appropriate for different learn-ing problems. Hence designing or applying discretization should not be blind to itslearning context. Section 6.5 provides a summary of this chapter.6.1 TerminologyDiscretization transforms one type of data[r]

of positive and negative examples in the i-th group respectively. At each stage, thefeature which minimizes Equation 5.1 is added to the current feature subset.The second algorithm chooses the most discriminating feature to add to the cur-rent subset at each stage of the search. For a given p[r]

fective for feature selection. Almuallim and Dietterich (1994) as well as Schlimmer(1993) have proposed forward feature selection procedure by constructing oblivi-ous decision trees. Langley and Sage (1994) suggested backward selection using thesame means. It has been shown that oblivi[r]

220 Richard A. BerkConsider now an application of the generalized additive model. For data de-scribed earlier, Figure 11.3 shows the relationship between number of homicidesand the number executions a year earlier, with state and year held constant. Indicatorvariables are included for[r]

. The signofβ3determines if the new line segment is steeper or flatter than the previous linesegment and where the new intercept falls.The process of fitting line segments to data is an example of “smoothing” a scatterplot, or applying a “smoother.” Smoothers have the goal of constructin[r]

possible hyperplane is the decision surface that assigns a new point to the class whosemean is closer to it. This decision surface is geometrically equivalent to computingthe class of a new point by checking the angle between two vectors - the vector con-necting the two cluster means and the[r]