Expert Systems Are Needed in Governmental Accounting and Auditing

Traditional expert systems have been developed and used in many aspects of accounting for several years. Expert systems have been constructed for auditing, tax and managerial accounting tasks. During the past two years, GAAP-based expert systems have been built and placed into use. This new type of system provides guidance in solving complex accounting problems involving leases, business combinations, etc. To date, little developmental work on governmental accounting expert systems has been reported. Yet, many issues in governmental accounting would lend themselves to expert system construction. Since the benefits of expert systems are documented, the potential for expert system development in the governmental accounting area appears high. The purpose of this article is to describe the components of an expert system and how they work. A brief description of traditional and GAAP-based expert systems that have been developed is provided. Finally, several examples of areas in governmental accounting suitable for expert system construction are discussed. EXPERT SYSTEM COMPONENTS The working parts of an expert system are a knowledge base and an inference engine. In addition, expert systems include a user interface program, and in the most powerful expert systems an explanation facility is included. Exhibit 1 depicts the relation ship among the components of an ex pert system. (Exhibit 1 omitted) KNOWLEDGE BASE The knowledge base includes facts about a specific topic or domain and rules for applying those facts. These facts and rules explain procedures or identify relationships. In addition to the facts and concepts about a particular domain, the knowledge base includes the heuristic rules an expert has developed through years of experience in solving a particular problem. These rules of thumb are known as heuristics. Software engineers or knowledge engineers who specialize in expert systems obtain information from domain experts and enter it into the knowledge base. Domain experts are people who are experts on the topic area for which the expert system is being designed. Many expert systems store the knowledge an expert provides in the form of production rules. These if then type rules represent condition and action knowledge. The "if' part of the rule specifies the condition and the "then" part indicates the appropriate action. For example if condition A is true then take action B. If the condition part of the rule is true then the action part is performed. An expert system that makes use of production rules is called a rule-based system. INFERENCE ENGINE Once the knowledge of an expert has been entered into the knowledge base, the inference engine enables the expert system to access the knowledge. The inference engine is a computer program which drives the expert system as it draws an inference by relating user supplied facts to a knowledge base rule and then proceeding to the next fact, rule combination. At each step, the inference engine recognizes the proper rule from the knowledge base and matches it with information supplied by the user. This process provides a chain in which the then part of one rule forms a link to the if part of the next rule. This chaining eventually leads to the most likely conclusion based upon the knowledge contained in the knowledge base and the inference drawn by the inference engine. USER INTERFACE The user interface is the component of an expert system that provides communication between the inference engine and the expert system user. The expert system asks questions and the user provides responses. The inference engine uses the user interface to make inquiries of the user for facts about the information contained in the knowledge base. In addition, the user interface delivers the expert system's conclusion and also explains the system's reasoning. It then relates the user supplied responses to knowledge base rules to make a decision about the problem. …