| Budget Amount *help |
¥1,280,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥180,000)
Fiscal Year 2007: ¥780,000 (Direct Cost: ¥600,000、Indirect Cost: ¥180,000)
Fiscal Year 2006: ¥500,000 (Direct Cost: ¥500,000)
|
|---|
| Research Abstract |
A technique for collecting shared data stored at low' cost on a complex organization, e.g., a company, a local government and a university, consists of several divisions is proposed to construct a novel web portal satisfying high browsability and retrievability. Usually, the shared data is collected from data producers by encouraging them to use a common format for a data management system. However, the burden of preparing the data in the common format causes the fail of collecting enough shared data. To tackle this problem, we have formulated a method for collecting the shared data automatically from raw data produced in a daily work at the organization. The collecting system involving the method is realized for the shared data existing in minutes of meeting in Oita University. The method automatically extracts the shared data for referring information about the meeting, i.e., names, date and time, meeting place, attendees, and absentees. The system allows a user in charge of writing the minute to just pass it to the system for providing shared data. The shared data extraction is done by using decision trees for judging whether each line of the minute includes the data for each of items. Attributes for describing any line are settled for each item. We have also designed interface components for facilitating checks and corrections of the extracted shared data by the user. The method showed high precision in the shared data extraction in the experiment using the minutes of past meetings of the university. The usability of the interface component was also verified in a user study we conducted. The component was shown to be less burdensome for the users in confirming the extracted shared data.
|
