Viability of Implementing Data Mining Algorithms as a Web Service


This paper describes an experiment into the viability of implementing data mining algorithms within a W3C standards compliant web service. The experiment shows that it can be done by the successful deployment of a prototype based on an implementation of the K-means clustering algorithm. The prototype produced demonstrates how the concept of a datamining web-service can be a reliable and effective data-mining tool especially in environments where raw processing power is a valuable commodity. The slim-client to fat-server model is demonstrated effectively showing how a user armed with a simple web browser can potentially harness super computing power. In addition the foundation for the development of an advanced datamining framework is presented which can include the implementation of any number of data mining techniques. The paper also seeks to highlight some ideas for future research and development of more sophisticate web services that are more scalable to suit both very specific tasks and very large datasets

Extracted Key Phrases

2 Figures and Tables

Cite this paper

@inproceedings{Stent2005ViabilityOI, title={Viability of Implementing Data Mining Algorithms as a Web Service}, author={Carl Stent and Nick Howard and Mohammad Saraee and Edward Thompson}, booktitle={ISWS}, year={2005} }