EXPERT SYSTEM
In general, expert systems are systems that try to adopt human knowledge to computer designed to model the ability to solve problems like an expert. With this expert system, ordinary people can solve the problem or just looking for a real quality information can be obtained only with the help of experts in the field. This expert system will also be able to assist the activities of the experts as an experienced assistant and has an experienced assistant and has the necessary knowledge. In composition, combining expert system rules of inference (inference rules) with a specific knowledge base provided by one or more experts in a particular field. The combination of these two things are stored in the computer, which then used in decision-making process for solving a particular problem.
Characteristics of Expert System
A good expert system must meet the characteristics as follows:
• Having reliable information.
• Easily modified.
• Can be used in various types of computers.
• Have the ability to learn to adapt.
Expert System Advantages
In general, many benefits that can be taken by the expert system, among others:
1. Allows ordinary people can do the work of experts.
2. Can do is repeat the process automatically.
3. Store of knowledge and expertise of experts.
4. Increasing output and productivity.
5. Improve quality.
6. Able to take and preserve the skills of experts (especially those including rare skill).
7. Able to operate in dangerous environments.
8. Have the ability to access knowledge.
9. Having reliability.
10. Increasing capabilities of computer system.
11. Have the ability to work with incomplete information and contains uncertainty.
12. As a complement in the media training.
13. Improving capabilities in solving problems.
14. Save time in decision making
Expert System Weaknesses
In addition to having several advantages, expert systems also have several weaknesses, among others:
1. Costs required to create and maintain very expensive.
2. Difficult to develop. This is of course closely related to the availability of experts in the field.
3. Expert systems are not 100% true value.
Expert System Structure
The main components in the structure of expert system according to Hu et al (1987) include:
1. Knowledge Base (Knowledge Base)
Knowledge base is the core of an expert system, namely the representation of expert knowledge. Knowledge base consists of facts and rules. The fact is the information about objects, events, or situations. This rule is a way to generate new facts from facts already known.
2. Machine Inference (Inference Engine)
Inference engine acts as the brain of the expert system. Inference engine serves to guide the reasoning process of a condition, based on the available knowledge base. Inside there is a process of inference engine to manipulate and direct the rules, models, and the fact stored in the knowledge base in order to reach a solution or conclusion. In the process, the inference engine using reasoning strategies and control strategies. Reasoning strategy consists of certain reasoning strategies (Exact Reasoning) and uncertain reasoning strategies (Inexact Reasoning). Exact reasoning will be done if all the data needed to draw a conclusion is available, while inexact reasoning done on state control sebaliknya.Strategi serve as a guide in conducting the reasoning procedure. There are three techniques frequently used controls, ie forward chaining, backward chaining, and the combination of these two control techniques.
3. Base Data (Data Base)
The database consists of all the necessary facts, which facts are the facts used to meet the conditions of the rules in the system. The database stores all the facts, whether the fact early on when the system began operating, and the facts obtained in the inference process is carried out. The database is used to store data observations and other data required for processing.
4. The user interface (User Interface)
This facility is used as an intermediary of communication between the computer user with
Knowledge Representation Techniques
Knowledge representation is a technique to represent the acquired knowledge base into a scheme / diagram that can identify specific relationships / connectedness between a data with other data. This technique helps knowledge engineers in understanding the structure of knowledge that will make the expert system. There are several knowledge representation techniques commonly used in the development of an expert system, namely
a. Rule-Based Knowledge
Knowledge is represented in a form of facts (facts) and rules (rules). This form of representation of premise and conclusion.
b. Frame-Based Knowledge
Knowledge is represented in a form of hierarchy or network frames.
c. Object-Based Knowledge
Knowledge is represented as a network of objects. Object is a data element consisting of data and methods (processes).
d. Case-Base Reasoning
Knowledge is represented in the form of conclusions cases (cases).
With Rule Inferencing: Forward and Backward Chaining
Inference with the rules is the implementation of the component mode, which is reflected in the search mechanism (search). Can also check all of the rule in the knowledge base in the direction forward or backward. Search process continues until there is no rule that can be used, or to a destination (goal) is reached. There are two methods by inferencing rules, the forward chaining or data-driven and backward chaining or goal-driven.
a. Backward chaining
• Using a goal-driven approach, starting from the desired expectations of what happens (hypothesis), then check with the causes that support (or contradictory) of these expectations.
• If an application produces a narrow, tree deep enough, then use backward chaining.
b. Forward chaining
• Forward chaining is a group of multiple inference that a search of a problem to the solution.
• If the premises in accordance with clause situation (TRUE), then the process will be to assert
the conclusion.
• Forward chaining is data-driven because the inference starts with the information available and obtained a new conclusion.
• If an application produces a wide tree and not in, then use the forward chaining.
In general, expert systems are systems that try to adopt human knowledge to computer designed to model the ability to solve problems like an expert. With this expert system, ordinary people can solve the problem or just looking for a real quality information can be obtained only with the help of experts in the field. This expert system will also be able to assist the activities of the experts as an experienced assistant and has an experienced assistant and has the necessary knowledge. In composition, combining expert system rules of inference (inference rules) with a specific knowledge base provided by one or more experts in a particular field. The combination of these two things are stored in the computer, which then used in decision-making process for solving a particular problem.
Characteristics of Expert System
A good expert system must meet the characteristics as follows:
• Having reliable information.
• Easily modified.
• Can be used in various types of computers.
• Have the ability to learn to adapt.
Expert System Advantages
In general, many benefits that can be taken by the expert system, among others:
1. Allows ordinary people can do the work of experts.
2. Can do is repeat the process automatically.
3. Store of knowledge and expertise of experts.
4. Increasing output and productivity.
5. Improve quality.
6. Able to take and preserve the skills of experts (especially those including rare skill).
7. Able to operate in dangerous environments.
8. Have the ability to access knowledge.
9. Having reliability.
10. Increasing capabilities of computer system.
11. Have the ability to work with incomplete information and contains uncertainty.
12. As a complement in the media training.
13. Improving capabilities in solving problems.
14. Save time in decision making
Expert System Weaknesses
In addition to having several advantages, expert systems also have several weaknesses, among others:
1. Costs required to create and maintain very expensive.
2. Difficult to develop. This is of course closely related to the availability of experts in the field.
3. Expert systems are not 100% true value.
Expert System Structure
The main components in the structure of expert system according to Hu et al (1987) include:
1. Knowledge Base (Knowledge Base)
Knowledge base is the core of an expert system, namely the representation of expert knowledge. Knowledge base consists of facts and rules. The fact is the information about objects, events, or situations. This rule is a way to generate new facts from facts already known.
2. Machine Inference (Inference Engine)
Inference engine acts as the brain of the expert system. Inference engine serves to guide the reasoning process of a condition, based on the available knowledge base. Inside there is a process of inference engine to manipulate and direct the rules, models, and the fact stored in the knowledge base in order to reach a solution or conclusion. In the process, the inference engine using reasoning strategies and control strategies. Reasoning strategy consists of certain reasoning strategies (Exact Reasoning) and uncertain reasoning strategies (Inexact Reasoning). Exact reasoning will be done if all the data needed to draw a conclusion is available, while inexact reasoning done on state control sebaliknya.Strategi serve as a guide in conducting the reasoning procedure. There are three techniques frequently used controls, ie forward chaining, backward chaining, and the combination of these two control techniques.
3. Base Data (Data Base)
The database consists of all the necessary facts, which facts are the facts used to meet the conditions of the rules in the system. The database stores all the facts, whether the fact early on when the system began operating, and the facts obtained in the inference process is carried out. The database is used to store data observations and other data required for processing.
4. The user interface (User Interface)
This facility is used as an intermediary of communication between the computer user with
Knowledge Representation Techniques
Knowledge representation is a technique to represent the acquired knowledge base into a scheme / diagram that can identify specific relationships / connectedness between a data with other data. This technique helps knowledge engineers in understanding the structure of knowledge that will make the expert system. There are several knowledge representation techniques commonly used in the development of an expert system, namely
a. Rule-Based Knowledge
Knowledge is represented in a form of facts (facts) and rules (rules). This form of representation of premise and conclusion.
b. Frame-Based Knowledge
Knowledge is represented in a form of hierarchy or network frames.
c. Object-Based Knowledge
Knowledge is represented as a network of objects. Object is a data element consisting of data and methods (processes).
d. Case-Base Reasoning
Knowledge is represented in the form of conclusions cases (cases).
With Rule Inferencing: Forward and Backward Chaining
Inference with the rules is the implementation of the component mode, which is reflected in the search mechanism (search). Can also check all of the rule in the knowledge base in the direction forward or backward. Search process continues until there is no rule that can be used, or to a destination (goal) is reached. There are two methods by inferencing rules, the forward chaining or data-driven and backward chaining or goal-driven.
a. Backward chaining
• Using a goal-driven approach, starting from the desired expectations of what happens (hypothesis), then check with the causes that support (or contradictory) of these expectations.
• If an application produces a narrow, tree deep enough, then use backward chaining.
b. Forward chaining
• Forward chaining is a group of multiple inference that a search of a problem to the solution.
• If the premises in accordance with clause situation (TRUE), then the process will be to assert
the conclusion.
• Forward chaining is data-driven because the inference starts with the information available and obtained a new conclusion.
• If an application produces a wide tree and not in, then use the forward chaining.
0 komentar:
Posting Komentar