Developing Kowledge Management Systems Step by Step.

Many organizations implemented knowledge management (KM) systems with little thought to deployment methodology, which led in part to the failure of 50 to 60 percent of all deployments (G.Lawton, IEEE Computer, 34(2), 2001). In this paper we study the methodology of developing KM applications. We classify general KM systems into three grades, i.e., Prime KM (PKM) systems, Standard KM (SKM) systems and Advanced KM (AKM) systems. A PKM system is the main and groundwork part of a general KM system for an organization. The basic functions of a PKM can be stated in the following three aspects: (1) keeping and organizing working experience and knowledge; (2) guiding employees in learning knowledge and (3) offering the basic intelligent services in supporting human creative activities. The intelligent services include offering the convenience in knowledge reuse and retrieval, linking various information and knowledge resources in Intranet/Internet, supporting knowledge distribution, visualizing knowledge flow and synergy relationships and providing knowledge filters for knowledge base updating. A SKM = A PKM + the functions of supporting know -how exchanging among members and partners. A so-called Know-how Exchange is utilized for implementing the functions, where the vertices of the network are members, divisions or partners and the arc between two vertices are established dynamically in logic based on one of relationships, e.g., the relationships of having common research interests, sharing of market information and cooperating in design and services. The network helps employees in finding “who”, “how”, “where” as well as support knowledge distribution. An AKM = A SKM + the ability of knowledge discovery in a special domain. We introduce our designs and implementations on the prototype systems of PKM and SKM in the domain of supporting research groups and projects respectively, and then we propose our design on a prototype system of AKM for marketing. The main techniques of knowledge discovery used in the design of the AKM are listed as follows: (1) the techniques of data mining and knowledge discovery Internet resources; (2) establishing the profiles and the purchase patterns of customers; (3) analyzing the dependency on the sale of two different kinds of merchandises and the relativity of two different kinds of purchase activities; (4) detecting the deviation of purchase patterns. We provide all key techniques on the development of the three prototype systems based on our research and implementations done by far, e.g.