Intelligent mining agent

Data mining is a complex, iterative and multistage process. These characteristics of data mining make it a complex field that needs the interference of professionals and experts with a lot of effort and work. In this work, a data mining system is proposed using a software agent that have the ability to handle the steps of mining automatically. The role of the agent is to fetch databases from the environment as a sensory input, pass the different stages of data mining, take decisions when necessary, and then stores the results in a knowledge base file as an output. This automation enables non-experts to use the system and facilitate the work of experts too. The task of the proposed system is divided into four stages: At first, the preprocessing stage in which some preprocessing techniques are applied if needed. Next, the task selection process stage in which the agent analyzes the attribute characteristics in the metadata file to select the appropriate mining task. Then, the mining stage for applying the algorithm and obtaining results; and finally the result representation stage for storing results in a knowledge base file.