Knowledge Management System
 
The RippleDown™ product is a knowledge acquisition and management tool that is most applicable for applications that have the following characteristics:
 
  • The application has a "decision support" component, that is, expertise is applied in a concrete situation in the form of some decision. The "decision" can take many forms, for example, it could be a textual report for a set of test results, or an insurance premium amount for a particular customer query, or a set of recommended actions for a set of alarms that may go off, etc.
  • The application area is "complex", that is, the decisions are based on several inputs, and may require several thousands of rules to be developed (over time) to cover all the types of decisions.
  • The application area is "evolving" rather than static, that is, the decisions may change as new knowledge or guidelines or user preferences chanage over time
  • The application is "high volume", that is, the expertise is required to be applied to very many case, and in particular, more frequently than can be practically managed by a human expert.
  • The application requires "high accuracy", that is, each decision must be the correct one, as judged by the expert, in the vast majority of cases (for example, greater than 99.9% of the time)
  • The application has "an expert in the loop" to review all (or a sample of) the decisions made, and to teach the system how to correct itself when it makes a mistake.
 
Example areas where RippleDown™ are being used, either commercially or in prototype form:
 
  • Providing reports from Proteomics or DNA test results
  • Providing reports from Environmental test results
  • Retrieving documents from the WEB according to user specifications - Intelligent Document Retrieval
  • Fraud detection, based on patterns of behaviour
  • Personal Information Management, for example, intelligent e-mail assistance
  • Data Mining of large data sets according to complex criteria