11.07.2015 Views

Encyclopedia of Computer Science and Technology

Encyclopedia of Computer Science and Technology

Encyclopedia of Computer Science and Technology

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

188 expert systemsAnatomy <strong>of</strong> an Expert SystemAn expert system has two main components, a knowledge base<strong>and</strong> an inference engine. The knowledge base consists <strong>of</strong> a set<strong>of</strong> assertions (facts) or <strong>of</strong> rules expressed as if . . . then statementsthat specify conditions that, if true, allow a particularinference to be drawn (see prolog). The inference engineaccepts new assertions or queries <strong>and</strong> tests them against thestored rules. Because satisfying one rule can create a conditionthat is to be tested by a subsequent rule, chains <strong>of</strong> reasoningcan be built up. If the reasoning is from initial factsto an ultimate conclusion, it is called forward chaining. If aconclusion is given <strong>and</strong> the goal is to prove that conclusion,there can be backward chaining from the conclusion to theassertions (similar to axioms in mathematical pro<strong>of</strong>s).While some rules are ironclad (for example, if a closedstraight figure has three sides, it’s a triangle) in many realworldapplications it is necessary to take a probabilisticapproach. For example, experience might suggest that if acustomer buys reference books there is a 40 percent chancethe customer will also buy a related CD-ROM product.Thus, rules can be given weights or confidence factors <strong>and</strong> asthe rules are chained, a cumulative probability for the conclusioncan be generated <strong>and</strong> some threshold probabilityfor asserting a conclusion can be specified. (See also fuzzylogic <strong>and</strong> uncertainty).While rules-based inference systems are relatively easyto traverse automatically, they may lack the flexibility tocodify the knowledge needed for complex activities (such asautomatic analysis <strong>of</strong> news stories). An alternative approachinvolves the construction <strong>of</strong> a knowledge base consisting<strong>of</strong> frames. A frame (also called a schema) is an encodeddescription <strong>of</strong> the characteristics <strong>and</strong> relationships <strong>of</strong> entities.For example, an expert system designed to analyzecourt cases might have frames that describe the roles <strong>and</strong>interests <strong>of</strong> the defendant, defense counsel, prosecutor, <strong>and</strong>so on, <strong>and</strong> other frames describing the trial <strong>and</strong> sentencingprocess. Using this knowledge, the system might be able topredict what sort <strong>of</strong> plea agreement a particular defendantmight reach with the state. While potentially more robustthan a rules-based system, a frames-based system faces thetwin challenges <strong>of</strong> building <strong>and</strong> maintaining a complex <strong>and</strong>open-ended knowledge base <strong>and</strong> <strong>of</strong> developing methods <strong>of</strong>reasoning more akin to generalized artificial intelligence(see artificial intelligence).Building an expert system requires that the knowledge <strong>of</strong> expertsbe “captured” in the form <strong>of</strong> a series <strong>of</strong> assertions <strong>and</strong> rules calleda knowledge base. Once the knowledge base is established, usersseeking advice can use an inference engine to examine the knowledgebase for valid conclusions that can be expressed as recommendations,<strong>of</strong>ten with varying degrees <strong>of</strong> confidence.TrendsExpert systems (particularly <strong>of</strong> the rules-based variety) nowhave an established place in business, industry, <strong>and</strong> science.The field <strong>of</strong> genomics <strong>and</strong> genetic engineering, widelyseen as the “technology <strong>of</strong> the 21st century” may be a particularlyfruitful applications area for analytical expert systems.Another promising area is the use <strong>of</strong> expert systemsfor e-commerce marketing analysis (see data mining). Anemerging emphasis in expert system development is theuse <strong>of</strong> object-oriented concepts (see object-oriented programming)<strong>and</strong> distributed database <strong>and</strong> knowledge sharingtechnology to build <strong>and</strong> maintain large knowledge basesmore efficiently.Further Reading“Expert Systems.” American Association for Artificial Intelligence.Available online. URL: http://www.aaai.org/AITopics/html/expert.html. Accessed July 31, 2007.Giarrartano, Joseph C., <strong>and</strong> Gary D. Riley. Expert Systems: Principles<strong>and</strong> Programming. 4th ed. Boston: Thomson Course<strong>Technology</strong>, 2004.“Introduction to Expert Systems.” Available online. URL: http://www.expertise2go.com/webesie/tutorials/ESIntro/. AccessedJuly 31, 2007.Jackson, Peter. Introduction to Expert Systems. 3rd ed. Reading,Mass.: Addison-Wesley, 1998.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!