Expert systems (Architecture)
-1 XLIFFService: Sequence contains no elements The architecture of expert systems, as a rule, consists of several components, presence of each of which ensures system operation in general. These components are important not so much separately how many in the harmonious interconnected work since all of them play an important role in solving of tasks for which the expert system is intended.
Content |
Enter into structure of expert system:
- solver
- working storage or database
- knowledge base (KB)
- components of acquisition of knowledge
- explanatory component
- dialogue component
Solver
The solver is intended in order that, using initial data from a working storage and knowledge from BZ to create such sequence of rules which, being applied to initial data, allow to solve a necessary problem.
Database
Initial and intermediate data of a solvable task are stored in the database (working storage). This term matches by the name, but not on sense the term used in information retrieval systems and the database management systems (DMS) for designation of all data stored in a system.
Knowledge base
The knowledge base is necessary for storage of the long-term data describing the considered area and the rules describing necessary actions over data of this area.
Components of acquisition of knowledge
Components of acquisition of knowledge automate processes of filling of ES knowledge, performed by the user expert. The expert of the high level enters the knowledge into expert system, more precisely, in a component of acquisition of knowledge. All heuristic method of solving of tasks is also based on this component in the subsequent.
Explanatory component
The explanatory component interprets in a form, available to the user, methods of the solution of a task or acceptance by the system of a certain solution. Besides, it performs functions of an explanation of an order of use of the data necessary for decision-making. It facilitates to the expert system testing and increases trust of the user to the received result.
Dialogue component
The dialogue component is necessary for the organization of friendly communication with the user, both during solving of tasks, and in the course of acquisition of knowledge and an explanation of results of work. The name of this component speaks for itself – it performs functions of the dialogue interface. With its help questions and amendments are entered into the system and also answers are visualized.
Work of expert system
ES can work in two modes: mode of acquisition of knowledge and mode of consultation. The consultation mode is called still the mode of the solution of a task or a usage mode of expert system.
Mode of acquisition of knowledge
In the mode of acquisition of knowledge through mediation of the knowledge engineer communication with ES is performed by the expert. In this mode the expert, using a component of acquisition of knowledge, fills a system with the knowledge. This knowledge, in turn, allows ES in a usage mode already independently to solve problems from a problem area. The expert describes a problem area in the form of a certain community of data and rules. Objects, their characteristics and values existing in a problem area are defined by data. Rules define the data management methods characteristic of the considered area. To the mode of acquisition of knowledge in a traditional development approach of programs there correspond the stages of algorithmization, programming and debugging executed by the programmer. Thus, unlike traditional approach in case of ES development of programs is performed not by the programmer, but the expert who is not owning programming.
Consultation mode
In the consultation mode communication with ES is performed by the end user whom the result and, perhaps, a method of its receiving interests. The user can not be a specialist in this problem area, depending on appointment of ES (in this case he addresses ES only for the end result), or to be a specialist (in this case he can receive result, but should address ES on purpose or accelerate process of obtaining result, or assign routine work to ES). In the consultation mode data on a task, after processing by their dialogue component, come to a working storage. A solver, being guided by the entering data from a working storage, the general data on a problem area and rules from the database formulates the solution of a task. ES at the solution of a task not only performs the ordered sequence of transaction, but also previously creates it. If reaction of a system is not clear to the user, then he can demand an explanation: "Why a system asks a question?", "as the answer is received?".
Representation of knowledge in expert systems
The main issue resolved at representation of knowledge is a question of determination of structure of knowledge. The second question concerns a form of representation of knowledge. Two of these problems are independent from each other, the selected method of representation can be unsuitable or inefficient for expression of some knowledge.
The question of formulation of knowledge can be separated into two rather independent tasks: how to organize knowledge and how to provide knowledge in the necessary form.
The necessity of selection of a formulation of knowledge in an independent task is caused, in particular, and methods of the solution of this task are that this task arises for any language of formalization identical or similar regardless of the used representation forms. The issues resolved at representation of knowledge are as follows:
- determination of structure of the represented knowledge
- organization of knowledge
- representation of knowledge
At determination of structure of knowledge the following factors are considered:
- problem environment
- architecture of expert system
- requirements and purposes of users
- communication language
Taking into account architecture of expert system of knowledge separate into interpreted and not interpreted. The first type: that knowledge which the solver is capable to interpret. All others treat the second type. The solver does not know either their structure, or contents. If these knowledge is used by any component of a system, he does not "realize" this knowledge.
Not interpreted knowledge
Not interpreted knowledge divides into the auxiliary knowledge storing information on lexicon and grammar of language of communication, on structure of dialog, and maintaining knowledge. Auxiliary knowledge is processed natural language komponenty, but the course of this processing does not realize a solver as this processing stage of entrance messages is auxiliary for conducting examination. The supporting knowledge is used during creation of a system and at an explanation of actions or solutions.
The interpreted knowledge
The interpreted knowledge can be separated into subject knowledge, managing knowledge and knowledge of representation. Knowledge of representation contains information on how in a system the interpreted knowledge is provided.
Subject knowledge contains data on data domain and methods of conversion of these data at the solution of assigned tasks. Let's note that in relation to subject knowledge of knowledge of representation and knowledge of management is metaknowledge. In subject knowledge it is possible to select descriptors and actually subject knowledge. Descriptors contain certain information on subject knowledge, such as coefficient of determinancy of rules and data, measures of importance and complexity. Actually subject knowledge breaks into the facts and the performed statements. The facts define possible values of entities and characteristics of data domain. The performed statements contain information on how it is possible to change the description of data domain during solving of tasks. Speaking in other words, the performed statements are the knowledge setting processing procedures.
Managing knowledge can be separated into focusing and solving. The focusing knowledge describes what knowledge should be used in this or that situation. Decisive knowledge contains information used for the choice of a method of interpretation of knowledge suitable to the current situation.
Connected subjects
Links
| The site content is translated by machine translation software powered by PROMT. The machine-translated articles are not always perfect and may contain errors in vocabulary, syntax or grammar. Read original article If you find inaccuracies or errors in the results of machine translation, please write to editor@tadviser.ru. We will make every effort to correct them as soon as possible. |