In this article I have shown how semantic nets add depth and power to business process models. In creating robust and extensible business models we should investigate ways to improve the capabilities and vigor of if-then rule based systems before turning to more advanced computational intelligence techniques such as evolutionary programming and adaptive neural networks. Rule based models still provide a lot of bang for the buck (as die conventional return on investment analysis goes). They are easy to understand, easy to modify, easy to tune, and easy to produce from historical data. Rule based models can also explain their reasoning — a real plus in todays world of fast moving, complex eBusinesses. When we connect our rule-based models to powerful semantic nets, we gready enhance their problem resolution capabilities. More to the point, semantic business models not only aid eBusinesses in a wide variety of customer preference analyses, new product positioning, and long term attrition analysis, but, by their very nature, support an easy to use diagnostic mechanism to look at the root cause of business problems. A complete semantic net construction and deployment library, with C source code, is available form the author. The code compiles into a dynamic link library as a set of Application Program Interface (API) calls that can be used from C++, Java, or Visual Basic. For previous readers of the author's articles who may have requested — but did not receive — the Java Expert System code or die Self-Organizing Map (SOM) code, I regret this problem (caused by a very punitive act of God). Please send another email to the author and these previous code examples will be shipped right away.
展开▼
