Vendor Implementations - Java Data Mining: Strategy, Standard, and Practice - page 422

Table 16-6 KXEN Java Data Mining supported functions, and function level tasks

Supported function Supported function level tasks

Classification BuildTask

ClassificationTestTask
ClassificationTestMetricsTask
DataSetApplyTask
RecordApplyTask
ComputeStatisticsTask
ExportTask

Regression BuildTask

RegressionTestTask
RegressionTestMetricsTask
DataSetApplyTask
RecordApplyTask
ComputeStatisticsTask
ExportTask

Attribute Importance BuildTask

ComputeStatisticsTask

Association BuildTask

ComputeStatisticsTask

Clustering BuildTask

DataSetApplyTask
RecordApplyTask
ComputeStatisticsTask
ExportTask

processed in KXEN as if “missing” was a specific category and then
encoded properly.

16.2.5 DME URI and Data URI Specifications

The only mechanism provided by KJDM to create a connection
toward a DME is through a specification provided as a URI. Java 1.4
includes an implementation of URI representation and parsing in the
package java.net that KXEN uses internally in KJDM. KXEN uses a
hierarchical URI structure of the form:

[scheme:][//authority][path][?query][#fragment].
The KXEN URI for DME connections accepts only two schemes:
kxjni to deal with in-process integration and kxcorba to deal with
out-of-process integration. The KXEN URI uses the query section of