1.1 Concept of Web Knowledge Flow
Many efforts have been done on the knowledge flow area. Some researchers study knowledge flow based on the organization of workflow, which is about the knowledge demand of the logic relationship and the role between the tasks of workflow. It can be realized by the way of pushing (Zhao, 2001; Wolverton, 1997). Taxonomy model aims at providing an overall picture of grid workflow verification and validation (Chen, 2007). Chen and Yang develop a novel checkpoint selection strategy that can adaptively select not only necessary but also sufficient checkpoints (Chen, 2007). Spiral model proposed by Nonaka describes the knowledge flow from epistemology to ontology. In the epistemology, knowledge flows from implicit knowledge to explicit knowledge, then from explicit knowledge to implicit knowledge. In the ontology, knowledge flows from person to group, and then from group to person (Nonaka, 1994; Nonaka, 1995). Based on Spiral model, Knowledge Flow Dynamic Model (KFDM) proposed by Nissen makes the knowledge flow over time explicitly. It can support a multi-dimensional representation that enables a new approach to analyze and visualize diverse knowledge flow patterns in enterprises (Nissen, 2002). Knowledge energy model proposed by Zhuge et al. takes knowledge energy as the driving cause to form an autonomous knowledge flow and explores the hidden principles (Zhuge,2005). The principles of knowledge flow engaged in cooperative cognition are explored by Dou from the perspective of learning and cognition evolution (Dou, 2006). Textual knowledge flow proposed by Luo et al. aims to provide an effective technique tool and theoretical support analysis for the discovery and cooperation of knowledge innovation, intelligent browsing, and personalized recommendation in Web services and e-Science Knowledge Grid (Zhuge, 2002). Other knowledge flow models are peer-to-peer team knowledge sharing and management based model (Luo, 2008), agent based model (Nissen, 2004), and the trust based model (Guo, 2005), etc.
Definition 1(Web Knowledge Flow, WKF):
Web Knowledge Flow(WKF) is a sequential link between topics with rich semantics, which is activated by user’s demands and changes with the demands.
Compared with other multiple types of knowledge flow (Nissen, 2002; Zhuge, 2005), WKF has some special characteristics as follows.
● WKF reflects the flow of knowledge between topics on Internet;
● WKF contains rich semantics between topics, which leads to similar WKF, associated WKF, and causal WKF etc;
● WKF is activated by user’s demands, in other words, WKF is a kind of service on demands;
● WKF changes with the change of user’s demands.
When a user browses topics, a WKF is a browsing path of topics recommended to the user. When some browsing paths of topics are activated at the same time, one of them should be chosen and recommended according to user’s demand.
As can be seen, Web knowledge flow provides a technique and theoretical support for the effective discovery of knowledge innovation, intelligent browsing, personalized recommendation, cooperative team work, and the semantic analysis of resources on Internet, which is a key issue of Web services and Knowledge Grid.
According to different relationship between nodes in Web Knowledge Flow, WKF can be classified into Association Web Knowledge Flow (Luo, 2008) and Similarity Web Knowledge Flow (Luo, 2008).
Definition 2:(Association Web Knowledge Flow, AWKF)
Associated knowledge flow is a sequential link with rich semantics between associated topics, which is activated by user’s demands and changes with the demands.
Compared with other knowledge flows, AWKF has the following distinguished characteristics.
● AWKF can reflect the flow of knowledge between topics;
● AWKF can contain associated relation between topics;
● AWKF can be activated by users’ demands, i.e., AWKF is a kind of service on-demand;
● AWKF can updates with users’ demands.
When a certain user browses topics in the Web or an e-Science environment, AWKF is a browsing path of topics. When several browsing paths of topics are activated at the same time, they should be evaluated according to the user’s demand and their contents, and then one of them should be chosen and recommended to the user.
Definition 3:(Similarity Web Knowledge Flow, SWKF)
Similar knowledge flow is a sequential link whose only difference from AWKF is that the relationship between topics in SKF is similarity relationship.
Therefore, if the user wants to browse similar topics on the Internet, an SWKF is helpful.
We know that web resources have two features in common:
● Various Types of Representation: Web resources exist in many kinds of representation: video, audio, text, and so on. First, it is impossible that the huge amounts of web resources are represented in one form. Second, different users have different requirement, which make the variety necessary. Even for one user, he/she must have different requirements in different time;
● Out-of-order Organization Statement: Web resources are distributed in out-of-order statement, which generates some challenges to the user’s web activity, such as web search. First, out-of-order statement makes resource searching cost lots of time, which affects the search quality. Secondly, out-of-order statement can not lead to high accuracy of search.
On the other hand, requirement of users is not limited to one form of media. For example, the user may want to browse the web resources whose contents meet his requirement, whatever its representation form is.
Therefore, in order to facilitate high qualified web activity it is indispensable to do the following work:
These are the basis for the generation of WKF. Element Fuzzy Cognitive Map(E-FCM) is used to represent web resource, whatever it is a video, audio, or text file. Semantic Link Network(SLN) is used to organize the web resources in order by introducing certain relationship. Next sections will include the contents of these two aspects.