Scientific Journal of KubSAU

Polythematic online scientific journal
of Kuban State Agrarian University
ISSN 1990-4665
AGRIS logo UlrichsWeb logo DOAJ logo

Name

Totukhov Konstantin Evgenyevich

Scholastic degree


Academic rank

—

Honorary rank

—

Organization, job position

Kuban State Technological University
   

Web site url

—

Email

ke.dnw@mail.ru


Articles count: 3

207 kb

ANALYSIS OF FACTORS AND INDICATORS OF THE INFLUENCE OF THE INTERNET ON THE INTELLIGENCE USING DATA MINING

abstract 1201606051 issue 120 pp. 755 – 769 30.06.2016 ru 410
In the article, we present the analysis and improvement of existing methodologies for assessing the intelligence factor, taking into account its use of the Internet on a regular basis. Today, the use of the Internet in our daily activities and tasks become practice that is more common. The Internet has become a popular and most frequently used means of obtaining information quickly and in large volume. The authors suggested the presence of the Internet and its impact on the everyday use of psychological and intellectual life of the people that can be recorded because of user IQ scores. It is suggested here also gets its development in the search for and analysis of system models, allowing determining the numerical values of the parameters. Specifically, it is proposed to use a large variety of statistical data to clarify the unknown parameters that determine the levels of human intelligence, taking into account the impact of the Internet. These statistics include information such as the number of people with higher education, the number of Internet users, the degree of the Internet penetration in society, the most typical tasks that use the Internet users, and others. For the processing and analysis of the test statistics in the work proposed to use tools data mining, ie, data mining. We have also considered the most typical approaches Data mining, applied in similar research areas. We have analyzed what specific principles and methods can best approach to solving the problems of intelligence assessment indicators. The result of the article was a number of conclusions, in particular, the feasibility of applying clustering to analyze data in the field. Also, in certain cases, we provided the use of Kohonen neural network in the vector quantization network format. Methods: analysis of scientific literature and online sources of information on the current level in the field of IQ research; modeling; Methods systematization (tabular calculations and compilation); Experiment (real human test data). Methodological bases of research: a systematic approach (considering IQ in the form of a functional multi-component dependencies), probabilistic and statistical approach (provides guidance for constructing mathematical models linking together IQ and the influence exerted on it using the Internet, and to assess the reliability of a computer program ) qualimetric approach (determines the need multicriterion diagnostic influence of factors on the level of intelligence)
252 kb

HYBRID NEURO-EXPERT SYSTEM FOR IDENTIFICATION OF SIGNIFICANT EVENTS TO SCHEDULE TIME SERIES

abstract 1241610049 issue 124 pp. 756 – 769 30.12.2016 ru 475
This article discloses the use of hybrid neural / expertnetwork systems to the problem of finding the significant events of these studies market behavior. The neural network is trained by back propagation, and is used to highlight trends over time. The expert system is used to determine the degree of significance of data
347 kb

PATTERN RECOGNITION IN THE CHART CONTROL BASED ON NEURAL NETWORKS WITH REINFORCEMENTS

abstract 1241610050 issue 124 pp. 770 – 789 30.12.2016 ru 442
This article discloses the use of neural networks to recognize patterns in control charts. To recognize unnatural situation under control is possible by analyzing the chart pattern. Neural networks with reinforcements are the third generation of neural networks. In this study they are available for recognition in the management chart patterns. The article also discusses options for improvement of the learning algorithm in the form of additional rules for the synoptic pauses, time constants and switching threshold neurons
ßíäåêñ.Ìåòðèêà