Finding Educational Resources on the Web: Exploiting Automatic Extraction of Metadata

Searching for educational resources on the web would be greatly facilitated by the availability of classificatory metadata, but most web educational resources put up by faculty members provide no such metadata. We explore using text classification and information extraction techniques to automatically gather such metadata. Text classifications orthogonal to topic matter appear possible with high ( > 90%) accuracy, and exact-match information extraction has an F measure around 50%.