MADAM : A Distributed Data Mining System Architecture Using Meta-Learning

Data mining is an important field which involves extraction of hidden knowledge in datasets that goes beyond simple analysis. It uses a plethora of machine learning algorithms to build models which help knowledge discovery. Meta-learning enhances learning and improves the performance of a data mining system by removing biasness of a learning algorithm used in a data context. Data mining systems face several issues due to large volumes of data which are often distributed. Agent paradigm offers exciting new ways to realize data mining systems and addresses issues of the system by leveraging on its benefits. In this paper, we present Multi-Agent Data mining Architecture using Meta-learning (MADAM) — a distributed data mining architecture that combines multi-agent system and meta-learning with an objective to enhance the performance of distributed data mining system