5
al. 2000). In Example 3, a semi-supervised algorithm is used for image recognition. Although they are
less common, semi-supervised algorithms are garnering acceptance by business practitioners.
Figure 2 lists some of the most common algorithms in supervised, unsupervised, and semi-supervised
learning. Tables in Appendix A list the algorithms that are used in each type of learning, show the names
of corresponding SAS procedures and SAS Enterprise Miner nodes, and provide references where you
can find more detailed information. For the few mainstay algorithms that are not currently implemented by
SAS, you can pursue custom solutions by using Base SAS
®
, SAS/STAT
®
, SAS/OR
®
, and SAS/IML
®
tools.
SAS integration with R and Java add even greater extensibility to the SAS platform. Integration with R is
available through SAS/IML and the Open Source Integration node in SAS Enterprise Miner. A Java API is
provided by the Base SAS Java object.
Figure 2. Machine Learning Taxonom y