C.B. Baruque and L.A. Rivera (Brazil)
Digital library, data warehousing, data mining, clustering, and learning objects.
The LO-DL (Learning Objects Digital Library) Project is being developed at PUC-Rio in the Database Tecnology Lab (TECBD). This Project aims at integrating LOs repositories through their metadata in a uniform catalog or Digital Libray (DL), making it transparent to the users their locations and characteristics. In [1] we proposed a new approach for developing DLs using Data Warehousing Architecture, which is a method that addresses the main issues of DL development. We make comparisons between the components and services of both the Data Warehousing and the “Digital Librarying” Architectures. Furthermore, we suggest the use of Data Mining Techniques in some steps of the DL building and utilization. One of them has to do with the process of generating Specific Libraries (Data Marts) from the Central Libraries (Data Warehouse). This process is executed by using the parameters chosen by the owner of the Library being developped, that are stored in a Table. Another way of executing such process is using the clustering technique, which is applied to the Central Library users’ accesses LOG. The resulting clusters substitute the parameters normally used to create the Specific Libraries (Data Marts). In this paper we formalized the use of the clustering technique in order to define the scope of digital libraries created for users communities, what improves the searches performance and the quality of the library.
Important Links:
Go Back
