Интеграция диагностической системы, основанной на знаниях, с базой прецедентов


Кобринский Б. А. Благосклонов Н. А.


The hypothesis of the presumptive diagnosis before laboratory confirmation is especially important in orphan (rare) hereditary diseases. It is possible to solve this problem using computer-based decision support systems based on knowledge. However, in medical practice, there are cases of an atypical clinical picture in patients with fuzzy manifestations of features. In such cases, it is possible to increase the diagnostic accuracy using a precedent approach. The concept of “synthetic precedent” is introduced, which is the result of the transformation of an atypical case into a synthesized description. The paper presents methods for constructing synthetic precedents of two types. The precedents of the first type are created as a result of extension with the fuzzy boundaries for ordinal variables. The precedents of the second type are received by softening the requirement for the number of necessary signs of a patient to match an atypical case from the precedent library. An approach to the creation of a hybrid system, including a traditional knowledge base and a precedent library, is proposed and demonstrated. The use of the hybrid system increases the accuracy of early diagnosis of orphan diseases in childhood.

Внешние ссылки

DOI: 10.1007/978-3-030-86855-0_20

Читать в Google Книгах (англ.): https://books.google.ru/books?id=_SJGEAAAQBAJ&pg=PA287

Scopus: Ссылка

Web of Science: https://www.webofscience.com/wos/woscc/full-record/WOS:000711937600020

Microsoft Academic: https://academic.microsoft.com/paper/3202169943/

ResearchGate: https://www.researchgate.net/publication/355050501_Knowledge-Based_Diagnostic_System_With_a_Precedent_Library

Ссылка при цитировании

Blagosklonov N., Gribova V., Kobrinskii B., Shalfeeva E. (2021) Knowledge-Based Diagnostic System With a Precedent Library. In: Kovalev S.M., Kuznetsov S.O., Panov A.I. (eds) Artificial Intelligence. RCAI 2021. Lecture Notes in Computer Science, vol 12948. Springer, Cham. Pages 289-302