Incorporation of a Agent Based approach for Integrated OLAP/OLAM Architecture

摘要

The current data mining tools is used to build knowledge based on a huge historical data. At present, businesses are facing with fast growing data that are very valuable in contributing knowledge. Knowledge should be updated regularly in order to ensure its quality and precision thus improve the decision making process. Data mining has shown great potential in extracting valuable knowledge from large databases. However, current data mining algorithms and tools are costly and several are too complex in their operations when dealing with large databases. In recent years, agents have become a popular paradigm in computing, because its autonomous, flexible and provides intelligence. Embedding agents in the current data mining processes and tools are believed to be able to solve the obstacle .Data warehouses are used for information processing, analytical processing, and data mining. An OLAP based data mining is referred to as OLAP mining, or on-line analytical mining (OLAM), which emphasizes the interactive and exploratory nature of OLAP mining. This paper focuses on an agent-based integrated OLAP/OLAM architecture that mainly focuses on data preprocessing. The aims is to provides an auto preprocessing a set of new data, which suite to data mining novice user. The proposed agent based framework consists of eight supporting agents: data cleaning agent, transformation agent, data reduction agent, data integration agent, proxy MDDB agent, input agent, GUI agent and dividing agent. Also we have three co-coordinating agents: DB API agent, MDDB agent and CUBE API agent. This paper is start by introducing the data mining process problem includes data preprocessing which agent can solve data mining problems. By applying agent in data preprocessing, a tool that intelligence yet flexible can be produced.