Text Mining Medline

SVM Classification

Oracle Text utilizes the ODM SVM algorithm for document classification. Classification is a supervised method, requiring classification categories and training documents.

In this demo, documents returned by a search can be used as an individual categorie's training documents or as subjects of classification based on the current set of categories.

There are four SVM operation buttons.

• Initialize SVM :: Removes all existing classifier category DB objects
• View/Edit SVM Catagories :: Reports the names and training set sizes of all existing classifier categories. Individual categories can also be dropped here.
• Add SVM Category :: Creates a classification category (named as specifies by user in text area). Uses all documents from the current search result as category training data. Nothing prevents the same document from being used in the training set for multiple categories.
• SVM Classification :: Classifies each document from the current search result with existing categories

In this demo, classification results are displayed as lists of documents matching a specific category with a score of >50 (on a scale of 0-100). Some documents may not match any class and will not be seen. Others may fit multiple categories, especially if those categories are related. In practice, documents from any individual category could be subsequently clustered to determine significant topics.