Abstract

Content

Introduction

1. Relevance of the topic

2. The purpose and tasks of the research

3. Prospective scientific novelty

4. The concept of ontology

5. The introduction of ontology in an expert system

Conclusions:

References

Introduction

The field of artificial intelligence has the history of its development for more than forty years. From the beginning, it had a number of complex tasks, and the problem of their solutions remains relevant today. One of the objectives of the subject area is the problem of organization of expert systems (ES).

Operation of ES involves the use of knowledge and manipulating them on the basis of base of heuristic rules that have been formulated by experts. ES are intended for analysis, making consultations and diagnoses. They are focused on the tasks that would require the involvement of experts. The problem of expert systems is in the need to continuously update the knowledge base, the failure of which could lead to negative consequences.

1. Relevance of the topic

Expert systems are topical at the moment, but there are problems when using them after a certain period of time. These problems lies in the fact that the information used by the system is aging, the volume of knowledge base is insufficient to meet the new challenges. Updating data with the expert is not always possible so it is actually to use the updated knowledge base with the use of the Internet, both in automatic mode, and with the participation of the user. The idea is to introduce the ontology into an expert system that will allow the system to gain knowledge from the Internet.

2. The purpose and tasks of the research

The purpose of the master's work is the introduction of ontology into an expert system. To solve this issue problem; it is necessary to follow these steps:

- To consider the application of ontology classification;

- To consider the methods of expert systems construction;

- To identify how you can automate the updating of an expert system the knowledge;

- To determine, with the help of what approaches and methods, you can search the information.

3. Prospective scientific novelty

The estimated novelty lies in the development of models and algorithms of interconnection the expert system with the methods and approaches of ontology for searching in the Internet.

4. The concept of ontology

According to the definition of T. Gruber, an ontology is a specification of the domain conceptualization. This is formal and declarative representation which includes the concepts vocabulary and the corresponding terms in the domain, as well as logical expressions (axioms) that describe a set of relations within concepts. To describe the relationships in the ontologies used. The whole arsenal of formal models and languages, developed in the field of artificial intelligence as the predicate calculus, production systems, semantic networks, frames, etc is used to describe relationships in ontologies. Thus, today a trendy term “ontology” is close in the meaning to the term “artificial intelligence” [1].

5. The introduction of ontology in an expert system

If to imagine an expert system in general, we'll get a program consisting of classes and objects that are specified in feature values. Every time when accesses the system, the user specifies certain object properties for those one wants to get recommendations. System does not always have the knowledge that the user would like to get, so you need to implement a mechanism that will store information about querying the ES, in cases where the knowledge base wouldn’t find the right solutions. When based on this definite information, the program should collect statistics of user requests. If the queries, or a certain part of the structure of queries for which there is no solution, have repeated, there is the reason - it makes sense to search for information on external resources. In this case, a request is forming to the ontology, with already - the structure properties formed according to statistics. With the help of ontological methods, we search for information on external resources. Once we have requested a search engine, analyzed the information and already had a final result, we are able to transfer information from the ontology into our system, thus updating the knowledge. It is desirable that the process of updating the knowledge base to occur in the dialogue with the user, as it is not always the information posted on the Internet, is correct. Also, it is desirable to search the knowledge and update occurred on user demand, without statistic. There may be occasions when searching for information on the structure of a given query fails, thus updating the knowledge will not be produced.

You can make the updating of knowledge of ES in the automatic mode, that is, updating the data at regular intervals, but we must not forget about the share of the risk, as the information stored in the knowledge base will not be accurate. It is also worth noting that such activity will lead to hardware and resource costs, which will, lead consequence, to the additional financial costs.

It is desirable that in the ES has been implemented with the security system that will protect against unauthorized access, thereby reducing the number of unwanted user queries, which would involve a knowledge updating.

We must understand that the ontology and the system must not contradict each other, that is, their structure should be similar.

Classes and instances of ontology are transferred to ES. As a rule they are usually reflected in the classes and instances of the programming language in which the system is implemented.

Axioms, built during the creation of ontologies, are partly reflected in the structure of the programming classes ES, and partly in the rules of EN. After performing each rule, which modifies, deletes or adds facts in the ontology is control of the contradictions that will protect it from getting into an inconsistent state.