EXPLANATION MODEL IN AN INTELLIGENT INFORMATION SYSTEM BASED ON THE CONCEPT OF KNOWLEDGE COHERENCE
| Autor: | I. Leshchynska, Serhii Chalyi, Volodymyr Oleksandrovich Leshchynskyi |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
Interpretation (logic)
рекомендації lcsh:T58.5-58.64 Process (engineering) Computer science узгодженість знань lcsh:Information technology Harmonization знання Data science Range (mathematics) Order (exchange) інтелектуальна система Information system General Materials Science Relevance (information retrieval) пояснення вимоги до пояснення Abstraction (linguistics) |
| Zdroj: | Вісник Національного технічного університету "ХПІ": Серія: Системний аналiз, управління та iнформацiйнi технологiї, Vol 0, Iss 1 (3), Pp 19-23 (2020) |
| ISSN: | 2410-2857 2079-0023 |
| Popis: | The subject of the research is the processes of constructing explanations of the results obtained in intelligent information systems. Explanations make transparent the process of creating the resulting recommendation, create the conditions for the user to create causal links between the result of the conclusion and the current problems for which the intelligent system was used. The aim is to develop an explanation model in an intelligent system with the ability to provide a consistent interpretation with the results of such a system, taking into account context-oriented requirements for user needs. To achieve this goal, the tasks of defining the requirements for explanation of the results of the information system, as well as developing a model of explanation based on the principles of knowledge coordination in order to obtain recommendations based on the facts, hypotheses, results. It is shown that when coordinating the interests of the user and the capabilities of the intelligent system, it is necessary to detail the knowledge taking into account the range of relevance of data and knowledge, as well as financial and technical and other features of the results. A model of explanation in an intelligent system based on the coordination of knowledge and data is proposed. The model contains many agreed facts, hypotheses and results. Reconciliation is performed for hypotheses that are subsets of other hypotheses, based on explanations of hypotheses through facts and other hypotheses, as well as on the basis of the correspondence between the obtained results and hypotheses. The model makes it possible to limit the use of implicit and inaccurate knowledge in the output only to their agreed subset. In practical terms, the use of the model is focused on constructing explanations taking into account the level of abstraction, the degree of detail of the description of the subject area, as well as taking into account the sel ected aspect of interpretation of the received recommendation. In general, the formation of explanations based on the harmonization of knowledge increases user confidence and creates conditions for the effective use of the received recommendations. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|