With the advent of cost effective storage systems and high speed network connectivity, the amount of data gathered by various transactional systems has increased manifold and processing the same has become a real challenge. A number of data mining systems have been developed for processing such huge data in a meaningful way. Most of these systems are stand-alone in nature except a few of them that facilitate a subset of their overall functionality for web-based usage. This is mainly due to the

more »

... hallenges involved in designing the architectural framework required for developing a web-enabled data mining system. Such a system is aimed at analysing the transactional data pertaining to various input domains employing some advanced graph mining techniques like subgraph matching, frequent subgraph discovery and graph visualization. In this paper, we present an innovative approach employing various Java Enterprise Edition (JEE) based technologies for developing a data mining system that operates in the web environment. This paper basically presents the architectural principles required to design a suitable framework for enabling the development of such a system and the implementation challenges faced in realising the same. The paper also discusses functional details of the system involving various graph mining techniques. A few general guidelines based on our understanding during the system implementation are also included in this paper for completeness.