Our research works focus on distributed knowledge sharing for digital ecosystem. We have the following contributions:

·        Development of a framework for collaborative knowledge sharing between participants of the ecosystem;

·        Propose  new strategy of sharing distributed knowledge for digital ecosystems: Virtual Web;

·        Propose to use data mining and KDD for digital ecosystems

·        Propose a formal concept structure for digital ecosystems

·        Development of the system and new algorithms of distributed knowledge sharing for digital ecosystem

Framework for collaborative knowledge sharing

The framework includes Query service, Knowledge engine, Ecosystem Sharing service  and P2P service, and knowledge sharing interface.

The core of the framework is Knowledge engine and Ecosystem Sharing service. Knowledge engine is for acquirement of knowledge that responds to where and how to find knowledge. Ecosystem Sharing service is for sharing knowledge in digital ecosystem. It focuses on the answer to how to share knowledge.

Knowledge engine provides methods of knowledge searching if the knowledge is already existed, and the algorithms of data mining if knowledge is implicit in data. When knowledge is acquired, we need Evaluation, Interpretation for knowledge.

Knowledge need to be trusted, accepted, stored, distributed, shared and updated in  digital ecosystem. Ecosystem Sharing service will provide the methods, models and strategies to share distributed knowledge.

Peer-to-peer (or abbreviated P2P) architecture is a type of network in which each workstation has equivalent capabilities and responsibilities. This differs from client/server architectures. Generally, P2P networks are used for sharing files, but a  P2P network can also mean Grid Computing. The infrastructure of P2P provides a novel distributed environment, computational model, and unprecedented opportunities for unlimited computing and storage resources. It's distinguished from conventional distributed computing by its focus on large-scale resource sharing, innovative applications, and, in some cases, high-performance orientation. Grids can be used as effective infrastructures for distributed high-performance computing and data processing. P2P environment provides high performance computing facilities and transparent access to them in spite of their remote location, different administrative domains and hardware and software heterogeneous characteristics. In this framework, P2P provide the support of communication, computation, data storage and data processing for distributed knowledge sharing in digital ecosystem.

Data mining and KDD for digital ecosystems

In the digital ecosystem, some  knowledge is existing, but some knowledge is previously unknown, implicit, hidden in large data. So we should extract the knowledge from  large data. The techniques of data mining (DM)  are widely used in research and application to look for relationships and knowledge that are implicit in large volumes of data and are interesting in the sense of impacting an organization's practice. Hence, we propose to use the techniques of data mining to extract knowledge for the digital ecosystem. We will discuss main issues of application of data mining for the digital ecosystem.

Data mining is the automated analysis of large volumes of data for extracting knowledge that are implicit in data and are interesting in the sense of impacting an organization's practice. The techniques of data mining are widely used in research and application. For example, data mining can be applied in biology, medicine, physics, and engineering. Data mining techniques can help companies  to provide better, customized services and support decision making.

Knowledge Discovery in Databases  (KDD) is a process of discovering non-trivial, previously unknown and potentially useful information from a huge collection of databases. Data mining is a step of KDD process, but also referred to KDD. The KDD process is an interactive process. Because it involves many steps with decisions made by users, different KDD process may have different steps. However, there are some basic steps in common. With years of research and developments, the computer-aided analysis such as data mining, machine learning, and statistics analysis of databases plays an increasingly important role in knowledge discovery and data analysis.

In the digital ecosystem, knowledge is core for distributed knowledge sharing. We need develop the following models according to the features of knowledge:

·        Knowledge Extraction and discovery

·        Knowledge Sharing

·        Knowledge Adapting and Transferring

·        Knowledge Identifying

·        Knowledge Evaluation

·        Knowledge Visualization and presentation etc.

Virtual Web

The core of the framework of distributed knowledge sharing is Knowledge engine and Ecosystem Sharing service. We need efficient, dynamic, flexible and scalable models and services in these two parts. We propose a new strategy: Virtual Web, to provide efficient, dynamic, flexible and scalable models and services for distributed knowledge sharing. The Virtual Web has the following features:

·        A new strategy of sharing distributed knowledge

·        Providing various services of sharing distributed knowledge

·        Dynamic

·        Self-adapting, self-organizing

·        Knowledge storage

·        Data warehouse: storage of metadata, data and other resource

·        Knowledge organization, classification and clustering

·        Engine of data mining and web mining

·        Web-based searching and extracting

Formal concept structure for digital ecosystems

We propose to use FCA as a tool for data analysis, information management and knowledge representation  in Digital ecosystem.

Formal Concept Analysis (FCA) provides a natural platform for data analysis and knowledge representation.  FCA is different from some of the traditional, statistical means of data analysis and knowledge representation because of its focus on human-centered approaches. From the formal concepts, we can analyze data such as revealing stronger association or relation between itemset and the set of their common objects, classifying objects, generating implications of attributes or knowledge rules, extracting the hierarchical relation among formal concepts, etc.  Concept lattice facilitate exploring, searching, recognizing, identifying, analyzing, visualizing, restructuring and memorizing  conceptual structures.

The core of FCA is concept lattice. Theoretical foundation of concept lattice founds on the mathematical lattice theory. Lattice is a popular mathematical structure for modeling conceptual hierarchies. Concept lattice is a method for deriving conceptual structures out of data. From concept lattice, we can study the relations between objects and attributes in a formal context, and how objects can be hierarchically grouped together according to their common attributes. Certain object subset and the set of their common attributes can represent each other, such duality sets form a formal concept, which the attribute subset is called intent and the object subset is called extent. Among the formal concepts, it exists an order relation, they form a complete lattice: concept lattice. Each node in the lattice is a concept and the corresponding graph (Hasse diagram) that is considered  as the generalization and specialization relationships between concepts. Such graphical structure represents directly and visually the relations of conceptual hierarchies. It allows us to  analyze and mine the complex data for such as classification, association rules mining, clustering, etc.

Furthermore, concept lattice  also provides an effective tool of knowledge visualization. Visualization plays an important role in data analysis. For example, Concept lattice can be used as an effective tool of visualization of biomedical data mining. Complex structures and sequencing patterns of genes and proteins are most effectively presented in graphs, trees, cubes, and chains by various kinds of visualization tools. Such visually appealing structures and patterns facilitate pattern understanding, knowledge discovery, and interactive data exploration.

Development of the system and new algorithms

We are developing the system for distributed knowledge sharing and new algorithms for extracting knowledge in digital ecosystems.

Other works

Work has started on Deliverable D4.2 - Distributed Accountability model for an Autopoietic P2P network .

 

Work has started on Deliverable D7.2 - Community Network technologies for Digital Business Ecosystems.

 
Summary of outputs

Huaiguo Fu, Formal Concept Analysis for Digital Ecosystem,  in proceeding of The Fifth International Conference on Machine Learning and Applications (ICMLA'06), 2006, Orlando, Florida, USA

Jennings, B., and Malone, P. 2006, Flexible Charging for Multi-provider Composed Services using a Federated, Two-phase Rating Process, in Proc. 2006 IEEE/IFIP Network Operations & Management Symposium (NOMS 2006);

Malone, P., Jennings, B. and Walsh, F. 2006, Accounting for Dynamically Composed Services in Digital Ecosystems, in Proc. 2006 Information Tecnology and Telecommunications Conference (IT\&T), Carlow, Ireland, October 2006;

Huaiguo Fu, Brendan Jennings, Paul Malone, Analysis and Representation of Biomedical data with Concept Lattice, in proceeding of The 1st IEEE/IES Conference on Digital Ecosystems and Technologies, 2007, Cairns Australia

Razavi, A. R., Malone, P., Moschoyiannis, S., Jennings, B., and Krause, P. J. 2007, A Distributed Transaction and Accounting Model for Digital Ecosystem Composed Systems, in Proc. 1st IEEE International Conference on Digital Ecosystems and Technologies (DEST 2007), Cairns, Australia, February 2007;

Presentations:

Huaiguo Fu, Data Mining and Concept Lattice for Digital Ecosystem, in Tampere Integration Workshop,  10th-11th Oct 2006

Huaiguo Fu, The Framework of Distributed Knowledge Sharing for Digital Ecosystem, in P2P Architecture Workshop, 27th-28th Nov 2006

 

Partner list: