DUOMENŲ SAUGYKLŲ IR AGENTINIŲ TECHNOLOGIJŲ PANAUDOJIMAS NUOTOLINIO MOKYMO SISTEMOS DUOMENIMS PARUOŠTI DUOMENŲ GAVYBAI
USING DATA WAREHOUSE AND AGENT TECHNOLOGIES TO PREPARE E.LEARNING DATA FOR DATA MINING

Summary/Abstract: The aim of this work is the application of data mining technologies to e.learning groupware system’s data. The software which is analyzed in this paper was developed by IBM Lotus Notes/Domino and characterizes non traditional database model. In such model data are stored as a single objects and this is a serious problem in a way of deploying data mining software. We suggest a new method in order to avoid problems such as document based model data collection, transformation, aggregation and filtering, which are considering on agent and data warehousing technologies. Methods for registering and processing Internet data are fairly new. During the design of infrastructure for information technologies not enough attention has been paid to collecting data suitable for data mining analysis. All Internet data flows are caused by various devices. However, every device has native data formats and uses different algorithms, so the first stage of data analysis, data collection, becomes much more complicated. In this paper we tried to present example of applying the data mining technology for e.learning data. Invoking created agent and created data collection it became possible to apply data mining technology methods to the e.learning system’s document data. Considering peculiarity of documental database, methodology allowing to transform in the real time data from documental database into data warehouse, invoking agent technologies. Interpreted results are useful from the department point of view. It helped to change necessary document databases and in future create more personalized site and better site layout. In addition to that the analysis of servers’ activity was performed and servers scheduling was changed to a conditional non - busy time. This let escape a big workload and increase the gama and irma servers’ efficiency. The proposed agent deals with document collections and can work in automatic and manual regime.

• Details
• Contents