Automated Selection of Interesting Medical Text Documents by the TEA Text Analyzer

摘要

This short paper briefly describes the experience in the automated selection of interesting medical text documents by the TEA text analyzer based on the naive Bayes classifier. Even if the used type of the classifier provides generally good results, physicians needed certain supporting functions to obtain really interesting medical text documents, for example, from resources like the Internet. The influence of the functions is summarized and discussed. In addition, some remaining problems are mentioned.