In expert systems, expert system shells are the software that contains an interface, an inference engine, and the formatted skeleton of a knowledge base. In essence, the shell of an expert system is an empty container that must be filled with the items of expert knowledge that the inference engine can process for users. Expert systems are computer applications that provide specific problem-solving help that a user may need to access to resolve, for example, a difficulty operating utility software. A knowledge engineer would use this shell to develop the knowledge base and customize it to meet the needs of their specific customer user base. It would be customized to take input from a user and interpret that information for the data warehouse and, by comparison, find corresponding information that can help guide the user to a solution.
man holding computer
Along with the control information that is deposited in a knowledge base, there are rules and attribute definitions that govern the delivery of information to users. The knowledge base is built from specialization statements that mimic the analysis process of a human expert looking for enough knowledge to arrive at a solution. Expert system shells must provide capabilities to support the knowledge engineer’s work in developing a knowledge base that can operate as an expert system in real time. In such an expert system, the basis may be the constant change of data through data deletions or additions as industrial systems, networks, hardware, and software systems change over time. This constant change in input from other management systems should not hamper the base’s ability to reason at the same expert level, regardless of the changes.
Expert system shells provide the skeleton for mimicking expert human reasoning in rule methods known as forward chaining and reverse chaining. Direct chaining in these shells allows you to get data from a user and use the rules of the inference engine to find more data related to that data until there is enough information to reach a conclusion. Since the initial data received is what drives the search, this method is called a data-driven method. An application that illustrates this direct chaining method can explore the possibilities of arranging components within a computer to arrive at the best component placement.
Reverse threading collects data only when it needs it when a knowledge base is queried in a query. Its goal is to find a value for C and it reasons backwards to find the value of A and B that concludes the target value of C. This method of reasoning from current data to earlier data that was the basis of the current data is called goal driven. method. An application illustrating the shell inference rules of an expert system might include a physician inputting a current set of symptoms to obtain background information about the same or similar symptoms in background information for a specific medical diagnostic expert system.
Inferred knowledge is obtained by examining existing facts to arrive at probable new information. This is the reasoning process that inhabits the inference engine in expert system shells. This process is what initiates forward or backward chaining in rule-based expert systems. The inference rules that are built by inference engines in expert system shells are made up of conditional “if” clauses and “then” clauses in rule declarations that facilitate step-by-step guidance. These steps may be in the areas of financial services, human resources, and mortgage loan management, among others, to try to discover rules of thumb as possible recommendations when a definitive answer is not possible.
