A Free Trial That Lets You Build Big!
Start building with 50+ products and up to 12 months usage for Elastic Compute Service
Sixth chapter expert System
Teaching Content: This chapter mainly introduces the definition, structure, characteristics and types of expert system, analyzes the rule-based expert system, the framework based expert system and the model based expert system, sums up the new expert system of cooperative and distributed, and introduces the design method and development tool of the expert system with examples.
Teaching emphases: The characteristics of expert system, the type of expert systems and the design of expert system.
Teaching Difficulty: The design of expert system.
Teaching Method: Classroom teaching is the main. Pay attention to the theory of artificial intelligence and the representation of knowledge, and ask questions to deepen students ' understanding of basic principles and concepts as well as development and design of expert system. To assist in the understanding of abstract concepts by using the relevant content in web-based courses.
Teaching Requirements: Focus on the basic concepts and design of expert system, Master rule-based, model-based, framework based expert system, understand some concepts and types of new expert system, general understanding of expert system development tools and evaluation methods.
6.1 Expert System Overview
Teaching Content: This section discusses some basic concepts of expert systems, introduces the definition, structure, characteristics and types of expert systems. This section is an important part of this chapter and is the basis for an in-depth study of the expert system.
Teaching Focus: The definition of expert system, the structure of expert systems, the general characteristics of expert system, the tasks and characteristics of various expert systems.
Teaching difficulties: The structure and construction steps of expert system.
Teaching methods: Mainly through classroom teaching, explain various basic concepts and system structure, inductive expert system's general characteristics, analysis of various types of expert systems tasks, characteristics and examples
Teaching Requirements: Focus on the definition and basic structure of expert system, grasp the characteristics of expert system, understand the type of expert system
Characteristics of 6.1.1 Expert system
Expert system is a large number of experts in a field of expertise in the knowledge and experience of intelligent system, can use the knowledge of human experts and problem-solving methods to deal with this area of problems. In short, expert system is a computer program system that simulates human experts to solve domain problems.
2. Expert System Features
Heuristic: Expert system can use expert knowledge and experience to make inference, judgment and decision.
Transparency: An expert system can explain its own reasoning process and answer the user's questions, so that users can understand the reasoning process and improve the trust of the expert system.
Flexibility: The expert system can continuously increase knowledge, modify the original knowledge, and constantly update.
3, the advantages of expert system
Specifically, the following eight areas are included:
(1) The expert system can work efficiently, accurately, thoughtfully, quickly and tirelessly.
(2) The expert system solves the actual problem without the influence of the surroundings, and it is impossible to omit forgetting.
(3) The expertise of experts can be limited by time and space to promote valuable and scarce expertise and experience.
(4) Expert system can promote the development of various fields.
(5) Expert system can bring together the knowledge and experience of many field experts and their ability to solve major problems.
(6) The level of military expert system is one of the important symbols of national defense modernization.
(7) The development and application of expert system has great economic and social benefits.
(8) The research expert system can promote the development of the whole science and technology.
The type of 6.1.2 expert system
1. Interpreting expert system
The task is to infer what might happen in the future through an analysis of past and present known conditions.
The characteristic data volume is very large, often inaccurate, has the error, does not have the complete explanation from the incomplete information, and can make certain assumptions to the data, the inference process may be very complex and very long
Examples of speech comprehension, image analysis, system monitoring, chemical structure analysis and signal interpretation.
2. Prediction Expert System
The task determines their meaning by analyzing and interpreting known information and data.
The characteristic system processing data changes with time, and may be inaccurate and incomplete, the system needs to have the dynamic model adapting time change
Examples include meteorological forecasts, military forecasts, population forecasts, traffic forecasts, economic forecasts and grain yield forecasts.
3. Diagnosis Expert System
The task infers the cause of an object's malfunctioning (i.e., failure) based on the observed situation (data).
Features can understand the characteristics of the components of the object or object being diagnosed and the relationship between them, can distinguish between one phenomenon and another phenomenon concealed, can provide users with measurement data, and from inaccurate information to get the correct diagnosis as possible
Examples are medical diagnostics, electronic mechanical and software fault diagnosis, and material failure diagnosis.
4. Design Expert system
Task to find a sequence or step of action that can reach a given target.
Characteristics from a variety of constraints to obtain the desired design, the system needs to retrieve a larger possible solution space, can be experimentally constructed possible design, easy to modify, can use the existing design to explain the current new design.
Example VAX computer structure design expert system.
5. Planning expert System
Task to find a sequence or step of action that can reach a given target.
The characteristics of the goal to be planned may be dynamic or static, the need for future action to make predictions, the problems involved may be complex.
Examples are military command and dispatch system, ropes robot planning expert system, automobile and train dispatching expert system and so on.
6. Monitoring Expert system
A task continually observes the behavior of a system, object, or process and compares the observed behavior with the behavior it should have in order to detect anomalies and issue alerts.
The characteristic system has the quick response ability, the alarm must have the very high accuracy, can process its input information dynamically.
Example of an expert system for the prediction of sticky insects.
7. Control expert System
The task adaptively manages the overall behavior of a controlled object or object to meet the expected requirements.
Characteristic control expert system has many functions, such as interpretation, prediction, diagnosis, planning and execution.
Examples are air traffic control, business management, autonomous robot control, combat management, production process control and quality control.
8. Debugging expert System
The task gives an idea and method of dealing with the object of failure.
Features at the same time with the planning, design, prediction and diagnosis of expert system functions.
Examples in this regard are relatively rare.
9. Teaching Expert System
The task of the teaching expert system is to teach and tutor the students according to their characteristics, weaknesses and basic knowledge, with the most appropriate teaching plan and method.
(1) Simultaneously has the function and so on diagnosis and debugging.
(2) has the good Man-machine interface.
Examples Macsyma symbolic integral and theorem proving systems, computer programming languages and physical intelligent computers assisted instruction systems, and deaf-mute language training expert system.
10. Repair Expert System
The task processes the failed object (System or device) to return to normal operation. The repair expert system has the function of diagnosis, debugging, planning and execution.
Examples of the ACI telephone and cable maintenance repair system in Bell Labs, USA.
In addition, decision expert system and consulting expert system are also available.
1. According to students ' characteristics, weaknesses and basic knowledge, the most appropriate teaching plan and teaching methods for students to teach and tutor the expert system is:
A Interpreting expert system B. Debugging expert system C. Monitoring expert system D. Teaching expert system
2. An expert system for finding an action sequence or step that can reach a given target is:
A Design Expert system B. Diagnostic expert system C. Predictive expert system D. Planning expert system
3. An expert system that can process a failed object (System or device) and return it to normal operation is:
A Repair expert system B. Diagnostic expert system C. Debugging expert system D. Planning expert system
4. An expert system that can infer what might happen in the future through an analysis of past and present known conditions is:
A Repair expert system B. Predictive expert system C. Debugging expert system D. Planning expert system
Structure and construction steps of 6.1.3 Expert system
1. Simplified structure of expert system
The structure of expert system refers to the construction method and organization form of each component of expert systems. Whether the system structure is chosen properly or not is closely related to the applicability and effectiveness of expert systems. The choice of what structure is most appropriate depends on the system's application environment and the characteristics of the task being performed.
Figure 6.1 represents a simplified structure diagram of an expert system.
Fig. 6.1 Structure diagram of the expert system for simplifying the structure of Tutu 6.2 ideal expert system
2, the structure of the ideal expert system
As shown in Figure 6.2. Because each expert system needs to complete the task and the characteristic to be different, its systematic structure also is different, generally only has the figure the partial module.
Interface is the medium of information exchange between human and system, which provides intuitive and convenient interactive means for users.
A blackboard is a database used to record control information, intermediate assumptions, and intermediate results used in a system's inference process. It includes plans, agendas and
3 parts of the solution.
The knowledge base consists of two parts. Part is the data information that is known to relate to the current problem, and the other part is the general knowledge and domain knowledge to be used for reasoning.
According to the control knowledge given by the system Builder, the scheduler selects an item from the agenda as the next action to be performed by the system. The executor executes the action selected by the scheduler by applying the information in the knowledge base and the records in the blackboard. The main function of the coordinator is to revise the results to maintain the consistency before and after the new data or new assumptions are obtained.
The function of the interpreter is to explain the behavior of the system to the user, including the correctness of the explanation and the reasons for the other candidate solutions.
Question: The methods of knowledge representation that have been learned are those.
3. The difference between general application and expert system
The former integrates the knowledge of problem solving into the procedure implicitly, while the latter
The problem solving knowledge in its application domain is composed of one entity,
That is the knowledge base. The processing of the knowledge base is separated from the knowledge base
Control strategy. More specifically, the general application organizes knowledge into level two: data-level and program-level; Most expert systems organize knowledge into three levels; data, knowledge base, and control.
1 definition of expert system.
2 What is the difference between an expert system program and a regular application?
4. The construction steps of the expert system
See figure 6.3, the general steps for establishing a system are as follows:
(1) Design the initial knowledge base, including:
(a) knowledge of the problem, that is, to identify the nature of the problem, such as what to solve the task is, how it is defined, can be decomposed into child problems or subtasks, it contains the typical data and so on.
(b) Knowledge conceptualization, that is, the key concepts and relationships needed to generalize knowledge representation, such as data types, known conditions (states) and objectives (state), proposed assumptions, and control strategies.
(c) Conceptual formalization, that is, determining the form of data structures used to organize knowledge, applying various knowledge representation methods in artificial intelligence to transform the key concepts, sub problems and information flow characteristics related to the conceptualization process into more formal representations, including hypothetical space, process model, and data characteristics.
(d) Formal regulation, i.e., the preparation of rules, the transformation of formalized knowledge into statements and procedures that can be executed by the computer in a programming language.
(e) Legalization of rules, namely, the confirmation that the rules of knowledge are reasonable and that the validity of the rules is examined.
(2) Development and test of prototype machine
After selecting the method of knowledge expression, it is possible to set up a subset of the experiments needed for the whole system, which includes the typical knowledge of the whole model and only involves simple enough task and inference process related to the experiment.
(3) Improvement and generalization of knowledge base
The knowledge base and the inference rule are improved repeatedly, and more perfect results are summed up. After a considerable amount of time (for example, months to two or three years), the system reaches the level of human experts within a certain range.
6.2 rule-based Expert System
Teaching Content: This section introduces a rule-based expert system.
Teaching emphases: The work model and structure based on rule expert system.
Teaching Difficulty: The working model based on rule expert system.
Teaching method: Classroom explanation.
Teaching Requirements: Master the principle of rule-based expert system.
1. The working model based on rule expert system
Rule-based expert system is a computer program, which uses a set of rules contained in the knowledge base to process the specific problem information (fact) in working memory and infer new information by inference machine. Its working model is shown in Figure 6.4.
Fig. 6.4 Working model based on rule-based expert system
The rule-based expert system does not need a precise match of human problem solving, but can provide a reasonable model for solving the problem of replication by computer.
Question: What kind of knowledge inference methods have you learned?
2, the structure based on the rule expert system
The complete structure of a rule-based expert system is shown in Fig. 6.5. Among them, knowledge base, inference machine and working memory are the core of this expert system. The main part of the system is the knowledge base and inference engine. Based on the reasoning system discussed so far, the knowledge base is composed of predicate calculus facts and the rules of the discussion topic. The inference engine consists of all the processes that manipulate the knowledge base to interpret the information requested by the user-such as digestion, forward chain, or reverse chain. The user interface may include a natural language processing system that allows the user to interact with the system in a limited natural language form. It is also possible to use a graphical interface interface with menus. Explain the inference structure that subsystem analysis is executed by the system and interpret it to the user.
Fig. 6.5 Structure based on rule-based expert system
6.3 Framework-based expert system
Teaching Content: This section introduces a framework based expert system.
Teaching emphases: goal-oriented programming and framework based design, structure and general design method based on framework expert system.
Teaching Difficulty: The structure based on framework expert system.
Teaching methods: Classroom teaching.
Teaching Requirements: Master the structure based on the framework expert system.
1. Goal-oriented programming and framework based design
The framework based expert system is based on the framework, which uses object-oriented programming technology, framework design and goal-oriented programming to share many features. When designing a framework based system, the designers of expert systems call the goal a framework.
2. Architecture based on framework expert system
A framework-based expert system is a computer program that uses a set of frameworks contained within a knowledge base to work memory specific
The problem information is processed and the new information is inferred from the inference machine.
It uses frameworks rather than rules to represent knowledge.
To illustrate some of the knowledge values in the design and presentation framework, let's consider the human frame structure shown in Figure 6.6.
Classes, subclasses, and examples (objects) are used to represent the organization of a framework-based system.
3. General design method based on framework expert system
The main design steps based on framework expert system are similar to rule-based expert systems. The main difference is in how to view and use knowledge, in the design of the framework based expert system, the whole problem and everything to imagine as woven things
After identifying things, looking for ways to organize these things, for any type of expert system, its design is highly interactive process.
Fig. 6.6 The human frame hierarchy
Thinking: 1. The composition, representation and inference of the framework in knowledge representation.
2. How can the two confusing terms "target" and "framework" be distinguished?
Thinking: The similarities and differences between the framework based expert system and the rule-based expert system are discussed.
Question: What is the difference between the framework-based expert system and the rule-based expert systems in regard to and use of knowledge?
6.4 Model-based expert system
Teaching Content: This section introduces a model-based expert system.
Teaching emphases: Model and Integration mode based on model expert system, the basic structure of neural network expert system.
Teaching Difficulty: The working principle of expert system based on neural network.
Teaching methods: Classroom teaching.
Teaching Requirements: Master the principle of the expert system based on the model.
1, based on the proposed model expert system
There are different opinions on the research content of artificial intelligence. There is a view that: artificial intelligence is a variety of qualitative models of the acquisition, expression and use of computational methods to study the knowledge. Based on this view, a model-based expert system is proposed.
The advantages of using various qualitative models to design expert systems are obvious.
In many models, the artificial neural network model is the most widely used.
2. Expert system based on neural network
The neural network model has the essential difference from the knowledge representation, the inference mechanism to the control mode, and the logic based psychological model in the present expert system.
3, three kinds of neural network model and expert system integration mode
(1) The Neural network support expert system is mainly based on the traditional expert system, supplemented by the related technology of neural network.
(2) The expert system supports the neural network based on the relevant technology of the neural network, establishes the expert systems in the corresponding field, and uses the related technology of the experts to complete the work of interpretation.
(3) The Synergetic neural network expert system is divided into several sub problems, aiming at the characteristics of each sub problem, choosing neural network or expert systems to realize it, and establishing a coupling relationship between the neural network and experts
4, the basic structure of neural network expert system
Automatic acquisition of module input, organization and storage of expert-provided learning examples, the selection of neural network structure, call the neural network learning algorithm, knowledge acquisition for the Knowledge base. When the new learning instance is input, the Knowledge acquisition module automatically obtains the new network weight distribution through the learning of the new instance, thus updating the knowledge base. As shown in Figure 6.7.
Fig. 6.7 The basic structure of neural network expert system
5. Discussion on several problems of neural network expert system
(1) The knowledge representation of neural networks is an implicit representation.
(2) The neural network realizes the knowledge automatic acquisition through the example learning.
(3) The inference of neural network is a forward nonlinear numerical computation process, and it is also a parallel inference mechanism, the output of neural network output nodes is a numerical value, so an interpreter is needed to explain the output mode.
(4) Several independent expert systems in the same knowledge field can be combined into a larger neural network expert system.
Question: Why the rule-based expert system can not be combined into large systems.
6.5 New expert system
Teaching content: The characteristics of the general new expert system, two kinds of new expert systems: Cooperative and distributed expert system.
Teaching emphases: Characteristics of new expert system, cooperative expert system, distributed expert system.
Teaching Difficulty: The connotation of new expert system features.
Teaching methods: Classroom teaching.
Teaching Requirements: Mastering the characteristics of the new expert system and distinguishing it from the general expert system.
Characteristics of 6.5.1 New expert system
1. Parallel and distributed processing
Based on various parallel algorithms, a variety of parallel inference and execution techniques are used to work in multiprocessor hardware environment, which has the function of distributed processing.
2. Multi-expert system work together
In this system, a number of expert systems collaborate.
3. High-level language and knowledge language description
The expert system can generate the expert systems automatically or partially automatically.
4. Self-learning function
The new expert system should provide advanced knowledge acquisition and learning function.
5, the introduction of a new inference mechanism
In the new expert system, in addition to deductive inference, there should be inductive reasoning, various non-standard logic reasoning, and various reasoning based on incomplete knowledge and fuzzy knowledge, etc.
6. Self-correcting and self-perfecting ability
In order to make mistakes, we must first have the ability to identify the wrong, in order to improve must first have the standard of identification.
7, Advanced Intelligent human-machine interface
Understanding natural language and realizing direct input and output of speech, text, graphics and images are the demands of intelligent intelligence.
6.5.2 Distributed Expert system
1, the main purpose: the function of an expert system after decomposition distributed to a number of processors to work in parallel, so as to improve the overall system processing efficiency.
2, environmental requirements: can work in a tightly coupled multiprocessor system environment, can also work in a loosely coupled computer network environment, so its overall structure to a large extent depends on its hardware environment.
3, design and implementation of distributed expert system, the problem to be solved:
The functional distribution allocates the functions or tasks of the decomposed system to each processing node in a reasonable and balanced way.
The knowledge distribution is divided into the processing nodes according to the function distribution,
Interface design between the parts of the interface design is to achieve the various parts of communication and synchronization is easy to carry out, in order to ensure the completion of the overall task of the premise, as far as possible to make the various parts of each other independent, some of the links between the less the better.
System structure depends on the environment and nature of application, on the other hand, it depends on the hardware environment.
The drive mode is available in several driving modes.
(1) Control drive when a module is required to work, direct control to the module, or it as a process directly call it, so that it immediately work.
(2) Data driven by a general system module function is based on a certain input, start the module to deal with, give the corresponding output.
(3) Demand-driven drive, also known as "purpose-driven", is a top-down drive. At the same time, according to the principle of data driven to the data (or other conditions) of the module to work, output the corresponding results and sent to their respective modules.
(4) Event-driven the module can be driven to work only if all events in the corresponding event collection of the module have occurred. Otherwise, as long as one of the events has not yet occurred, the module will wait, even if the module's input data is all available.
6.5.3 co-expert system
The field of problem solving in general expert system is very narrow the application of single expert system is very limited and it is difficult to obtain satisfactory application.
Cooperative multi-Expert system is an important way to overcome the limitation of general expert system.
The cooperative multi-expert system can also be called "group expert System", which represents an expert system to solve a wider domain problem by integrating several different aspects of the similar field or one domain.
The system emphasizes the cooperation among subsystems, but does not focus on the distribution of processing and the distribution of knowledge.
Thinking: The difference from distributed expert system?
2, the design and establishment of a cooperative multi-expert system, the problem to be solved:
(1) The decomposition of the task
According to the domain knowledge, the total task is decomposed into several sub tasks, which are accomplished by several expert systems respectively.
(2) The export of public knowledge
The common part of the knowledge needed to solve each task is separated to form a common knowledge base, which can be shared by all the sub-expert systems. The specific knowledge of each task is stored in the specialized knowledge base of each sub expert system.
(3) Discussion mode
At present, many authors advocate using "blackboard" as the "garden" for the discussion of various systems. In order to ensure the consistency of data or information in the blackboard in a multiuser environment, some means of managing the database are needed to manage it and use it, so the blackboard is sometimes called "intermediate database".
(4) Adjudication issues
The solution to this problem is often very dependent on the nature of the problem itself.
(5) Drive mode
This problem is consistent with the corresponding issues to be considered in the distributed database. Although the cooperative multi-expert system and subsystems may work on a processor, there are still some ways to activate the subsystems according to the general requirements, that is, the so-called driving mode problem.
6.6 Expert System Design
Teaching Content: This section takes the design of a rule-based maintenance consulting system as an example to illustrate the design process of the expert system.
Teaching emphases: Describe the expert knowledge and application knowledge, explain the decision.
Teaching Difficulty: The representation and decision of expert system knowledge.
Teaching method: Classroom Teaching, example explanation.
Teaching Request: Through the example to let the student understand the expert system more deeply, grasps the expert system design technique preliminarily.
Description of 6.6.1 Expert knowledge
According to expert expression of knowledge, in the system design process mainly use the following 3 components: Hypothesis or conclusion, observation or observation, inference or decision rules. In expert, there is a strict distinction between observation and hypothesis. Observation is observed or measured, its value can be "true (T)", "False (F)", number or "Do not know" and other forms. Hypotheses are possible conclusions obtained by systematic inference. It is generally assumed that a measure of uncertainty is attached. An inference or decision rule is expressed as a production rule 1, a representation of a conclusion
The conclusions set out the scope of the expertise involved. In expert, each assumption is represented by shorthand mnemonics and formal explanatory statements written in natural language (Chinese, English, or other language the designer wishes to use). A mnemonic symbol is used to refer to assumptions when writing decision rules.
Example: The problem of car repair is expressed by a table.
2. Representation of observations
Observation is the observation or measurement result required to obtain a conclusion. They can usually be expressed in logical values: True (T), False (F) or "unknown", or numerically. Organizing the problem into a menu is a very effective method.
3. The representation of inference rules
The production rule is the most commonly used representation of decision rules, which can be divided into 3 categories according to the logical relationship between observation and hypothesis:
(1) Rules from observations to observations (FF rules)
The FF rules specify the true values of observations that can be deduced directly from the observed observations. Because by combining observations and assumptions, you can describe a more powerful production rule form.
(2) from observational to hypothetical rules (FH rules)
In many expert systems for classification, production rules can measure the degree of credibility of the production conclusion.
(3) from assumptions to hypothetical rules (HH rules)
HH (from hypothetical to hypothetical) rules are used to prescribe inference between hypotheses.
The use and decision interpretation of 6.6.2 knowledge
It is not a precise science to establish an expert system. Experts often provide a great deal of information and must try to extract the key elements of the expert's reasoning process and express them as accurately and concisely as possible.
1. Classification and selection of conclusions
According to the Order of evaluation, dividing the rules into rank and selection rules is the basic part of the control strategy in the process of inference. The order of the rules can be ranked according to the opinions of the experts. At the same time, it is necessary to study the impact of the rules ' evaluation order. The arrangement of the Order of the rules evaluation should make the same conclusion regardless of the order taken.
The application of the trusted metric in the production rule can not only reflect the uncertainty that exists in the expert knowledge, but also reduce the number of production rules.
2. The strategy of asking questions
It is difficult to give a best strategy for asking questions, and the quality of questioning depends to a large extent on whether the problem is clearly organized in advance. One of the key issues in a good query strategy is to include as many structures as possible. The problem should be divided into groups according to the common theme. A simple rule such as the FF rule allows you to enforce branching by topic in the questionnaire. If the information required for system inference is not accepted at the same time, you can have the following two kinds of questioning strategies:
(1) In some situations, experts collect the required knowledge in a predetermined sequence or in a fixed order.
(2) The system does not inquire in a fixed order, but makes some choices according to the specific circumstances.
3. Explanation of decision
The system's designers and users need the system to interpret the decisions it makes. But they have different requirements for decision interpretation.
(1) Explain to the system designer.
(2) Interpretation of the user of the system.
One method of interpretation is to use statements to illustrate the conclusion. The assumptions used by the system may be any form of statements containing instructions and suggestions. Sometimes the designers of a system can advance some explanations that are appropriate for a given hypothesis.
Question: If all observations can be obtained at the same time, and the study is only a classification problem, then how to apply a simple control strategy.
For example, in the case of car repair, you can give a general explanation of how much is explanatory, rather than rigidly dividing the conclusions into two categories: diagnosis and processing. Such statements can be in the following form: "As the cylinder of the car is flooded, the door pedal is trampled on or waited for 10 minutes." ”
6.7 Expert system Development tools
Teaching Content: This section introduces four kinds of main expert system development tools.
Teaching Focus: Skeleton-type tools (also known as Shell), language tools, construction aids and supporting environment.
Teaching difficulties: Language tools and supporting environment.
Teaching method: Classroom explanation.
Teaching Requirements: Understand the common development tools of expert system, master the application and support environment of language development tools.
1. Skeleton Type development tool
The expert system usually has the inference machine and the Knowledge Base two parts, but the rule set exists in the knowledge storehouse. In an ideal expert system, the inference machine is completely independent of the problem solving domain. The perfection or change of the system function depends only on the perfection and change of the rule set. Thus, using the expert system developed before, the rule of description domain knowledge is "dug out" from the original system, and only the inference machine part which is independent of the knowledge of the problem domain is retained, so the tool which is formed is called skeleton type tool. This kind of tool because its control strategy is predetermined, the use is very convenient, the user only must express the specific domain knowledge to be some rules to be possible.
Because the main skeleton of the program is fixed, in addition to the rules, users can not change anything, so skeleton tools have some problems to be solved, affecting its wide application.
2. Language-Type development tools
Language tools provide users with the basic mechanisms needed to establish expert systems, and their control strategies are not fixed in one or several forms, and users can influence their control strategies through certain means. Therefore, the structure of the language tool has a wide range of changes, indicating flexibility, the scope of adaptation is much wider than the skeleton tools.
3. Construction of auxiliary tools
System construction aids are composed of a number of program modules, some programs can help to obtain and express the knowledge of domain experts, some programs can help to design the structure of the expert system is being constructed. It is mainly divided into two types, one is the design of auxiliary tools, the other is knowledge acquisition aids.
4, supporting the environment
Support facilities are tools that help with programming, and are often used as part of the Knowledge engineering language. The tool-supporting environment is only an accompanying package to make the user interface more user-friendly. It includes four typical components: debugging aids, input and output facilities, interpretation facilities, and knowledge Base editor.
Based on the production system, this chapter first studies the basic problems of expert system, including the definition, type, characteristic, structure and construction steps of expert system. Then the expert system based on the different technology is discussed, that is, the second section is based on the rule expert system, the third section is based on the framework expert system and the fourth section is based on the model expert system. From the working principle and model of these systems, it can be seen that various techniques and methods of artificial intelligence are well combined and applied in expert system, which provides a good example for the development of artificial intelligence.
Some new thinking and new technology of computer science also play an important role in the development of expert system. The new expert system in the fifth chapter of this chapter is the result of the distributed processing and cooperative working mechanism in the application of computer science, which are distributed expert system and cooperative expert system respectively.
The sixth section of this chapter introduces the design of expert system, takes a rule-based maintenance consulting system As an example, explains the design process of expert system, and uses expert development tool to design. This will have a more specific and in-depth understanding of the expert system.
In order to improve the development efficiency, quality and automation level of expert system, the development tools of expert system are needed. The seventh section of this chapter introduces 4 major development tools, namely skeleton tools, language tools, construction aids, and supporting environments.
Expert system is one of the earliest and most effective fields of artificial intelligence application research. People expect it to have new development and new breakthroughs.
Start building with 50+ products and up to 12 months usage for Elastic Compute Service