The unique combination of features provided by Knowledge Builder include the following:
Graphical Knowledge: rules are implemented and displayed in a graphical format to allow easier updating of knowledge by a non IT user. In addition the rules can be broken down into linked or chained decision trees for easier organization and displayed in different formats such as decision tables and rule sets.
Fuzzy & Crisp rules: rules can be implemented in a combination of fuzzy or crisp logic that permits the introduction of uncertainty or relative probability into different diagnostic fault conditions.
Mined rules from data:in addition to rules derived from expert knowledge or heuristics rules can also be derived from historical operating data via in built data mining algorithms using rule induction. Mined rules can be converted to predictive models via a fuzzification option.
Genetic Algorithms:the genetic algorithm optimization option provides the capability to solve complex resourcing and scheduling problems within the Knowledge Builder environment and can be combined with predictive models of performance from data mining.
External DLLs:if process models or calculations are available in the form of DLL’s then these can be called from Knowledge Builder and referenced in diagnostic logic
Deployment options:there is a range of deployment options depending on user requirements including the capability of being called as a server to another process for integration with SCADA systems and Plant Data Historians.