This chapter illustrated various JDM concepts related to data specifi-
cations, classification, regression, attribute importance, association
rules, and clustering functions. We saw that data specifications are
divided into physical and logical specifications to facilitate reusabil-
ity. Model build settings provide function-specific settings and algo-
rithm settings that are used to tune the models for problem-specific
requirements. JDM provides test metrics for supervised models to
understand model quality. JDM supports model apply for super-
vised and clustering models, providing control over the output val-
ues. JDM defines algorithm settings for decision tree, support vector
machine, naïve bayes, feed forward neural networks, and k-means
algorithm settings. In the next chapter we explore how these con-
cepts are mapped to Java classes and interfaces in JDM.
2000] M. P. Brown, W. N. Grundy, D. Lin, N. Cristianini,
C. W. Sugnet, T. S. Furey, M. Ares Jr., D. Haussler, “Knowledge-Based
Analysis of Microarray Gene Expression Data by Using Support Vector
Machines,” Proc Natl Acad Sci USA , 97:262-7.
[Cluster Tutorial 2006] See http://www.elet.polimi.it/upload/matteucc/
[Cristianini/Shawe-Taylor 2000] Nello Cristianini, John Shawe-Taylor, An
Introduction to Support Vector Machines and Other Kernel-based Learning
Methods , Cambridge, UK: Cambridge University Press, 2000.
[DeBlasio 2001] Agnes DeBlasio, “Data Mining Application Helps BB&T
Increase Cross-Sell Ratio, Bank Systems & Technology,” http://www.
[DM Methods Poll 2006] “Latest KDnuggets Poll Results on Usage of
Data Mining Methods,” http://www.kdnuggets.com/polls/2006/data_
[Dragoon 2005] Alice Dragoon, “How to Do Customer Segmentation
Right?” CIO Magazine , October 2005, http://www.cio.com/archive/
[Han/Kamber 2006] Jiawei Han, Micheline Kamber, Data Mining, Second
Edition: Concepts and Techniques , San Francisco, Morgan Kaufmann,