The paper considers the issue of creating training sets and their scaling in machine learning problems. The subject of the study is the process of generating training sets based on examples in order to augment them.To implement the idea of expansion, it is proposed to use the transformation of existing examples of situations. The transformation of examples is based on a well-known optimization method -the method of coordinate descent.The paper describes the statement of the problem of transformations of example situations in terms of the introduced representation model. There are proposed algorithms that make it possible to obtain an extended set from the initial set of example situations specified using formal representations, which will include situations that meet the similarity criteria with these examples.The paper presents the testing of the proposed algorithms for expanding a set of example situations, carried out in order to form a data set for the studying artificial neural networks. The obtained results are of practical importance for training artificial neural networks used in intelligent decision support systems. The proposed algorithms make it possible to automate the formation of datasets using the available prepared and approved examples of typical situations and solving the transformation problem as the problem of finding the optimum of the similarity objective function.
Метод вывода решений на прецедентах многими авторами рассматривается как основа для создания систем интеллектуальных поддержки принятия решений в различных предметных областях. Знания в виде прецедентов <ситуация, решение> хранятся в системе и используются при возникновении некоторой новой проблемной ситуации. Для применения в актуальной ситуации верного решения предусмотрен механизм поиска такой ситуации, которая отвечает заданному критерию сходства с актуальной, и вывода пользователям того решения, которое образует прецедент вместе с ситуацией из базы. Один из важных комплексов задач CBR-систем связан с проблемой адаптации решений, которая возникает в тех случаях, когда при возникновении новой ситуации CBR-система не находит в своей базе прецедентов (БП) похожей ситуации и не может рекомендовать готового и надежного решения. Данная статья направлена на изучение вопросов адаптации решений. В работе поставлена задача адаптации решений и предложена алгоритмизация адаптации решений. В результате исследования задача адаптации разделена на два типа: поиск подходящего решения в цепочках программ действий в БП и сборка (синтез) нового решения. В первом случае предполагается использование некоторой части готовой программы действий в качестве нового решения для актуальной ситуации. Во втором случае из разных программ действий, хранящихся в БП, поэлементно собирается новое решение. Представлены алгоритмы адаптации решений. Предложенные задачи и алгоритмы позволяют находить решение при возникновении ситуаций, которые не описаны в базе прецедентов CBR-системы и, таким образом, повышают надежность ее работы. The case-based reasoning method is considered by many authors as the basis for creating intelligent decision support systems in various subject areas. Knowledge in the form of precedents <situation, solution> is stored in the system and used when some new problem situation arises. To apply the correct solution in an actual situation, a mechanism is provided for searching for such a situation that meets the specified criterion of similarity with the current one, and displaying to users the solution that forms a precedent together with the situation from the database. One of the important tasks of CBR-systems is related to the problem of adapting solutions, which arises when, when a new situation arises, the CBR-system does not find a similar situation in its precedents database (PD) and cannot recommend a ready-made and reliable solution. This article is aimed at studying the issues of adapting solutions. The problem of adaptation of solutions is posed in the work and an algorithmization of adaptation of solutions is proposed. As a result of the study, the task of adaptation is divided into two types: the search for a suitable solution in the chains of action programs in the PD and the assembly (synthesis) of a new solution. In the first case, it is supposed to use some part of the prepared action program as a new solution for the current situation. In the second case, a new solution is assembled element by element from different action programs stored in the PD. Solution adaptation algorithms are presented. The proposed tasks and algorithms make it possible to find a solution in the event of situations that are not described in the case base of the CBR-system and, thus, increase the reliability of its operation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.