BIZON is an innovative approach to dynamic value constellation modelling and governance for e-business. The main goal is to design and build a knowledge founded framework (consisting of ontologies, knowledge bases, semantic Web, Web data mining, machine learning) able to support and optimise a business environment characterised by product personalisation, demand anticipation and process selforganisation.
BIZON is conceived for use in value constellations with the following features:
- Business actors that consist of a multiplicity of sellers, buyers, producers, each of them with their own distinct features;
-- Actors exchange goods (products and services) which can be material or immaterial, but in any case contain a high degree of embedded knowledge. The sectors targeted as examples in this proposal will be the goods and services of the publishing and knowledge management industries;
- The goods considered are complex, and can be generated as variations and combinations of simpler goods. Goods production (e.g. variation and combination) and delivery requires planning of processes under economic and time constraints. Process planning requires co-ordination among actors (sellers and producers, but also buyers);
- Actor behaviour is simultaneously co-operative (to generate value) and competitive (to divide the generated value);
- The value constellation is dynamic in all its components. As time proceeds, actors can leave the constellation and new actors can enter it; the behaviour of each actor is subject to evolution; goods evolve as well, and follow their own life cycle, sometimes very short.
In the context of the features outlined, BIZON models the value constellation by means of a formal ontology. The ontology is built around the fundamental concept of value.
The e-business model describes the value lifecycle: e.g. who is offering what to whom and expects what in return; how the value is offered or requested by whom; what value is generated; what
value is perceived by whom and why; when and where the value is generated, perceived and exchanged; how it is generated, and so on.
This e-value ontology describes production and exchange processes, in the broadest sense. It encompasses marketing aspects (value perception has a lot to do with Web data mining), production aspects (value generation implies planning and scheduling of production processes) and other aspects, e.g. legal ones. Important concepts to be treated in the e-value ontology are time, planning and scheduling, products variation and combination, products and services personalisation, idiosyncratic features of buyers, anticipation of market trends.
The BIZON components such as:
-the BIZON text analysis, categorisation and profiling modules;
-Knowledge Management Description Language;
-BIZON enhanced Content Management System;
-BIZON enhanced Corporate Portal are built as interoperable and platform independent software packages, developed in java.
Data is stored with standard database systems, and managed and transmitted in XML format. Documents of any format can be stored and managed, and users can insert and modify them through a simple HTML interface. Machine Learning is strongly biased with the background knowledge offered by the semantic meaning of data being processed, annotations given for the data, but also high-level features that can be inferred by the ontology-based inference engine. All this information is fed as input to Machine Learning algorithms for classification, clustering and profiling that are the basic modules of the content-management system and corporate portal applications. The machine learning results consist of a number of independent modules capable of implementing the basic functionalities related to user/document representation, document categorization, user/document clustering and user profiling. These functionalities are easily adapted for different applications related to e-business.
Funding SchemeCSC - Cost-sharing contracts