13.6. Expert Systems

Expert systems (ES) are designed to emulate the human ability to make decisions in specific contexts, and have had a large impact in the world of AI. This is because expert systems are often concerned with decision making under uncertainty which is a highly valuable skill. An expert system, by means of a reasoning, can perform at a level comparable to or better than a human expert does within a specified area.  The goal of ES is therefore to solve complex problems by following a line of reasoning that is derived from human knowledge. This reasoning is normally represented by if–then–else statements, instead of conventional procedural code.

Expert systems are based on two main components: a knowledge base and an inference engine.

Knowledge base Is the set of rules that have been extracted from human knowledge in a specific area.
Inference engine Applies the rules from the knowledge base to known facts, in order to deduce new facts. Inference can follow a forward strategy, starting with the available data and using inference rules to extract more data (known as forward chaining), or proceed inversely (backward chaining). More advanced systems also include explanation capabilities, which can motivate particular decisions made by the system.

Expert systems have been reliably used in the business world to gain competitive advantages and forecast market conditions. In a time where every decision made in the business world is critical for success; the assistance provided from an expert system can be an essential and highly reliable resource.

Advantages to Expert Systems

Responses to decisions, procedures, and tasks that are repeated frequently can be accomplished by expert systems. Providing that a system’s rule base remains unchanged and regardless how many times comparable challenges are tested; the findings will remain consistent. Expert systems can explain why a result was reached and why it views one option over another to be the most reasonable. Expert systems do not have the restrictions that humans do and can work around them. The systems continue long after the human with the expertise is gone.

Disadvantages to Expert Systems

As all judgments are dependent on inference rules set up in the system, expert systems lack the common sense that may be required in some decision-making applications. In unique scenarios, the systems are also unable to respond creatively, and innovatively as human experts would. Expert systems are not always free from error and errors may emerge during processing as a result of logical faults in the knowledge base, resulting in wrong solutions.

Expert Systems in Fraud Detection

Expert systems can be used in fraud detection. A company can establish rules for detecting fraud and then allow the system to identify fraudulent behaviour based on the set of rules created. For example, a rule can be created to look for transactions that are originating from a particular location (one that may been known for criminal activity) and are over a certain amount. Since expert systems are based on rules that are programmed, it may be easy for cybercriminals to circumvent the system.  However, expert systems can still be useful as part of a larger fraud defense strategy that may also incorporate more intelligent systems [1]

  1. Lu, C. (2017, February 19). How AI is helping detect fraud and fight criminals. Retrieved December 3, 2021, from https://venturebeat.com/2017/02/18/how-ai-is-helping-detect-fraud-and-fight-criminals/.


Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Information Systems for Business and Beyond by Shauna Roch; James Fowler; Barbara Smith; and David Bourgeois is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book