开放存取
##reader.subscriptionAccessGranted##
订阅存取
卷 46, 编号 5 (2019)
- 年: 2019
- 文章: 8
- URL: https://journals.rcsi.science/0147-6882/issue/view/10759
Article
Time Series Anomaly Searching Based on DBSCAN Ensembles
摘要
This article suggests a technique for building an ensemble based on the DBSCAN algorithm. This technique uses the internal structure of a time series for adaptively selecting input parameters. When used in experiments, it shows a narrower variance and higher levels of anomaly detection using real and synthetic data compared with a number of popular approaches.
299-305
Semantic Technologies for Semantic Applications. Part 1. Basic Components of Semantic Technologies
摘要
This paper discusses the basic aspects of the modern understanding of semantic computations, semantic technologies, and semantic applications in the field of artificial intelligence. The basic terminology accepted in the work is introduced and specific examples of semantic applications, including industrial-level ones, are given. The paper demonstrates that the basic components of semantic technologies of artificial intelligence are ontologies and semantic models of their use, semantic resources, and the semantic component of the technology. The semantic resources contain information about the semantics of words and other entities, as well as means of refinement of these semantics. The semantic component is used to create formal descriptions of the meanings of natural language entities and numerically evaluate their pairwise semantic similarity. The available semantic resources are discussed and a comparative analysis of them is given. Information on natural language entity types (primitives) is given and then used for the practical purposes of building models of formal description of the meaning of texts in various semantic applications. The latter components of description of text semantics constitute the contents of the second part of this paper.
306-313
Classification of Text Documents Based on a Probabilistic Topic Model
摘要
An approach to text document classification that utilizes a probabilistic topic model, which is characterized by the fact that its training document set contains objects of only one class, is proposed. This approach makes it possible to identify positive samples (samples resembling the target class) in collections and streams of text documents. This article considers models created for solving the problems of text document classification and trained on samples of a single class, describes their key features. The Positive Example Based Learning-TM classification model is presented and a software prototype that implements it as a basis for classification of text documents is developed. Despite having no information about negative document samples, the model demonstrates a high level of classification accuracy that exceeds the performance of alternative approaches. The superiority of the Positive Example Based Learning-TM model with respect to the classification accuracy criterion when using a small training set is experimentally proven.
314-320
Certainty Factor Triunity in Medical Diagnostics Tasks
摘要
This paper suggests approaches to investigating and solving the problem of three factors that characterize the measure of expert confidence in the occurrence of symptoms in diseases, the timing of the manifestation of symptoms, and the frequency of symptoms in progressive hereditary diseases in five age groups that differ in clinical manifestations (a polyvariant character space). Linguistic scales of fuzzy characteristics (interval age and the occurrence of signs) and certainty factors should contribute to a more subtle and accurate evaluation of diagnostically significant traits and increase the effectiveness of diagnosis at different ages. The measure of confidence is determined with respect to each characteristic used for a given nosological form. In the process of assessing risk factors, specific features of the thinking of experts are considered, that is, intuition, confidence in their knowledge, and reflexivity (regarding emerging hypotheses). Various stages and variants of group expertise with the participation of a cognitive scientist are considered.
321-327
A Multi-Criteria Decision-Making Procedure with an Inherited Set of Starting Points of Local Optimization of the Scalarizing Functions
摘要
Abstract—This paper proposes a new interactive iterative procedure of searching for a preferred solution of a complicated non-linear multi-criteria optimization problem, in which global optimization of the scalarizing functions of criteria is too difficult because of numerous local extrema of the function and other reasons. In the suggested procedure, instead of global optimization of the scalarizing function of criteria, a large number of local optimization problems are solved on each iteration, while the set of starting points of local optimization processes is generated in a small neighborhood of the decision inherited from the previous iteration and the type of the scalarizing function of varies from iteration to iteration. The proposed procedure, named the Inherited Decisions Method, was applied for multi-criteria selection of rules for controlling the Angara cascade of reservoirs, while the rules are described by hundreds of parameters and the problem is characterized by more than two dozen decision criteria.
328-336
The Possibilities for Intelligent Analysis of Scientific Texts by Construction of their Cognitive Models
摘要
This paper provides a study of cognitive structures of scientific texts within the framework of the activity paradigm. We define mental actions as actions that are aimed at constructing new or reconstructing already existing abstract objects. As is typical for scientific communication, forms of implementing mental actions mediated by certain verbal units and textual techniques are considered as mental operations. This pilot study demonstrated the fundamental possibility of constructing a model of the intelligent structure of a scientific text, including an assessment of the textual representation of categories reflecting various mental operations, and the identification of its intelligent scheme. A formal presentation of mental operations used to “describe a phenomenon new to science” is proposed.
337-344
Initial Generation of Concepts in Image Processing Based on the Algebra of Fourier-Dual Operations
摘要
Abstract—Inductive concept generation in image processing on neural networks that perform the algebra of Fourier-dual model-determining operations is considered. A model of initial (i.e., when there are no earlier concepts) concept generation is proposed. The biological inspiration of the model is established. The phenomenon of cognitive saturation is proposed, which is expressed by stopping the growth of the performance evaluation of the formation of an inductive hypothesis with an increasing number of examples. The mechanism of this phenomenon is analyzed and the relationship with the information capacity of patterns of the internal representations of perceived images, due to the inherent characteristics of the cognitive agent, is given. The simulation results are given based on the example of the Darii syllogism inversion when the weights of links are recorded in plane and in volume.
345-355
Modeling of Collective Decisions by a Virtual Council
摘要
Abstract—The paper presents the results of modeling collective decision making in small groups for the situation of diagnosing arterial hypertension, namely: analysis of the known modeling methods and integration of knowledge and diagnostic decision support tools, development of the architecture of the Virtual Council research prototype and its software implementation, as well as experiments with the council.
356-365