Reference no: EM131680149
Please respond to the following discussions
In your own words, evaluate the architecture of rule-based Expert Systems (ES) and explain how the knowledge engineering process used to build an ES. Make a suggestion for an ES that would be useful in your place of work.
Most Expert Systems (ES) are composed of six components, according to Sharda et al. (2015):
• Knowldge acquisition
• Knowledge base
• Inference Engine
• User Interface (UI)
• Blackboard
• Justifier
Some ES include a knowledge-refining system as well, but this is not so common. A knowledge-refining system, when present, enables the ES to learn from its successes and mistakes to improve its future decisions.
The "Justifier", "Blackboard", and UI of an ES will be explained first. The "Justifier" is an explanation system which explains why the ES chose certain options and not others. This is one of the most unique functions of an ES because it differentiates ES from conventional computer systems that do not explain how they reach their conclusions.
The "Blackboard" is a memory space reserved by the ES for evaluating hypotheses and potential courses of action. The data associated with these are recorded in the Blackboard and evaluated recursively until an adequate conclusion has been reached. The UI of an ES usually processes natural language questions from users and returns answers in a question-and-answer format.
The "inference engine", knowledge base, and knowledge acquisition system will be described now. The "inference engine" is the programmed methodology used by the ES to form conclusions based on the knowledge base. This engine may also known as the "control structure" or "rule interpreter", depending on the ES.
The knowledge base for an ES contains the facts and heuristic rules which the inference engine consumes to produce conclusions. The "knowledge aquisition system" is dedicated to loading expertise from experts and documentation into the knowledge base of the ES such that it can be used by the inference engine. The knowledge acquisition system is essentially an Extract Transform Load (ETL) system (cf. data warehousing), but it is specific to ES construction and maintenance and is usually driven by a dedicated "knowledge engineer".
According to Sharda, et al. (2015), XCON is a rule-based Expert System developed by Digital Equipment Corp (later bought by Compaq) which provided customers with optimal systems configuration in approximately 1/20 to 1/30 of the time it took humans to accomplish this task (p. 482).
I work in the technical support department at Bomgar Corp., and a significant portion of the work our customer support reps do is configure Bomgar appliance settings to meet the needs of customers who are not familiar with the complexities and nuances of the Bomgar product configuration options. An online expert system delivered through the Bomgar Self Service Center which understood Bomgar product configuration settings would massively reduce the workload of Bomgar customer support reps and increase the time to resolution of Bomgar customer support contacts.
Another important task performed by Bomgar Support reps is identifying software bugs reported by customers and delivering patches for these bugs to customers. An expert system which understood the backlog of reported Bomgar bugs could efficiently convey known fixes to Bomgar customers and triage new bugs for humans on the Bomgar support team to investigate.
The BMC Remedy tool called "HelpDeskIQ" reportedly sorts incoming customer email such that each email is placed in the queue of the customer support rep who is most qualified to resolve the problem at hand (Sharda, et al., 2015, p. 483). Bomgar Support currently employs a team of dedicated "Customer Support Analysts" (CSA) who do this task manually, escalating complex problems to higher level technical support reps and rerouting customers to the New Sales Implementation team, as necessary.
An automated system such as HelpDeskIQ could significantly reduce the workload on the CSA team and accelerate the speed with which incoming communications are routed to the right contacts in Support, especially after normal business hours when humans are asleep or partying wildly.
References
Sharda, R., Delen, D., Turban, E., Aronson, J. E., Liang, T., & King, D. (2015). Business Intelligence and Analytics: Systems for Decision Support (10th ed.). Upper Saddle River, NJ: Pearson
D2
Knowledge is information that is meaningful in intellectual forms such as understanding, awareness, and ability. It is typically acquired by experience, information consumption, experimentation and thought processes such as imagination and critical thinking.
There are many types of knowledge proposed and still in debate to agree. Some of the types of knowledge: A Priori, Posteriori, Explicit, Tacit, Propositional, and Non-propositional.
Explicit knowledge:
Explicit knowledge can be defined in words, numbers, shared information from the data, scientific formulas, and so on.
Tacit knowledge:
Tacit knowledge is highly personal and hard to formalize, difficult to communicate or share. The different dimensions of Tacit knowledge: technical, know-how, and cognitive dimensions.
D3
Q3 - Expert Systems
I think rule-based ES includes knowledge acquisition, knowledge base, inference engine, user interface, blackboard, explanation and knowledge refinement.
But knowledge engineering process is the similar, it includes knowledge acquisition, knowledge representation, knowledge validation, interfacing and explanation and justification, so I think knowledge engineering process is used in the whole process of building up ES or knowledge engineering process is the foundation of ED establish? I would like to hear more from other people.
One example I would think about is CPOE (Computerized Physician Order Entry). It has a knowledge base and it helps physician to make better decisions about prescriptions and exams such as lab and radiology tests.
For example, when a physician plans to order some antibiotic for a patient's pneumonia, if the system noticed that this patient has allergic history to Penicillin, it will warn the physician about this allergic history, so physician will order some other type of antibiotic and will avoid the adverse incident.