04-24-2013 09:02 PM
Could someone tell me what kinds of algorithms are used with the SAS Content Categorization? I do not want to see the exact formula, just the names of the types of algorithms used such as K-nearest neighbor, decision trees, naive Bayes, greedy search, etc. Thanks in advance.
06-20-2013 12:41 PM
There are a variety of algorithms that have been used in SAS Content Categorization, many of which are proprietary implementations. In general, SAS Content Categorization uses algorithms from finite state automata (FSA) while others are based on graph theory leveraging various searching approaches (e.g., depth first search). Other more widely known algorithms include using frequent phrase extraction and maximum entropy classification.
Do you have any specific parts of SAS Content Categorization that you are interested in knowing more about with respect to algorithms in use? I can help in this sense by narrowing in to a smaller scope.