| NetMODEL's rationale is founded on the observation that more and more commerce-related transactions are becoming digital. This is happening not just on the supplier side of the value chain, but increasingly on the consumer side. That is, new digital interactive devices are emerging in the mass-market arena. These include the Internet, Digital Interactive Television, Personal Digital Assistants (PDA) and mobile broadband telephony devices. These devices have at least one important facet in common; all the marketing channel interactions between digital service provider (information content, product and service provision) and the consumer can be recorded and stored in digital databases. Thus, for the first time, aspects of consumer behaviour are available for analysis. The major problem -- the scale of analysis -- can be addressed by using data mining technology, which is designed to discover knowledge in large databases; in this case, consumer behavioural patterns. However, existing data mining algorithms -- artificial intelligence, database, or statistics based -- have several generic drawbacks. Firstly, the output of data mining algorithms is usually in a form that is not readily understandable to those most in need to the results - marketing professionals. In the context of electronic commerce, this hinders the application of discovered marketing intelligence, which is the basis for personalisation activities. Secondly, the results are not represented in a format that allows the exchange of discovered knowledge. Thus, the application or exploitation of data mining results is a tedious, often impossible task. Some existing attempts, mainly the Knowledge Interchange Format (KIF), have proven too generic to be applicable in the area of knowledge discovery. It is expected that much of future research in the data mining community will focus less on algorithmic improvements and more on the generic representation and application of discovered patterns. The outlined problem is a huge research area in itself, and NetMODEL will focus on sequential patterns to be applied in electronic commerce scenarios. NetMODEL will investigate and develop new ways of representing information generated from data mining algorithms. Specifically NetMODEL will focus on the development of techniques for representing the output of the MiDAS (Mining Internet Data for Associative Sequences) algorithm into a machine-readable and portable form. The NetMODEL representation will be based on the eXtended Markup Language (XML), and will act as a bridge between mining and visualisation. Aims The overall aim of NetMODEL is to produce a test-bed system capable of representing and manipulating knowledge from different sources, including the output of the MiDAS sequence detection algorithm, information about a web site, and information about site transactions The research method will validate how XML (utilising Resource Description Framework - RDF) may be used to represent this knowledge from heterogeneous sources, and facilitate the application of that knowledge in real-time via an Internet server under HTTP. | |
Add a project to the database.
Update the entry for this project.