СРАВНЕНИЕ МЕТОДОВ ИСКУССТВЕННОЙ ГЕНЕРАЦИИ ДАННЫХ ДЛЯ ГЛУБОКОГО ОБУЧЕНИЯ СИСТЕМЫ МОНИТОРИНГА
This article deals with the problem of input data generating for the creation and training of an artificial neural network, which is the basis of the classification module of a dynamic monitoring system of the manufacture performance indexes. The input data that was used to train the neural network was divided into the following categories: real data, generated data for a given distribution, and data obtained using the simulation approach. The simulation model was created using the apparatus of Petri nets. Further, for the data used in the work, classification rules were set, after which the artificial neural network was trained on each data set. At the next step, real data was submitted to the monitoring system, which are previously did not appear in the training and validation of neural networks. The final step of this study was to compare the results of the classification of the described approaches of artificial generation of values of enterprise input parameters with respect to the control data set.
Nested Petri nets (NP-nets) is an extension of Petri net formalism within the “nets-within-nets” approach, when tokens in a marking are Petri nets wich have autonomous behavior and synchronize with the system net. The formalism of NP-nets allows modeling multilevel multiagent systems with dynamic structure in a natural way. Currently there is no tool support for NP-nets simulation and analysis. The paper proposes translation of NP-nets into colored Petri nets and using CPN Tools as a virtual machine for NP-nets modeling, simulation and automatic verification.
Modern criminalists do not share a common opinion regarding the choice of parameters which could be used to work out a system of characteristics to differentiate a maniac killer from an ordinary person. This hinders the development of efficient software for investigation purposes. The paper describes the experience of developing a neural network that can learn using data about known serial killers, including their biological, social and psychological parameters. The authors also evaluate the errors of this neural network’s mathematical model, prove its adequacy and present a study which allowed to work out a comparative evaluation of the impact that different parameters have on the modeling result, i.e. a person’s predisposition to violence. They prove that the most important parameters include: a psychiatric disorder, alcoholic parents, growing up with parents, family status, social status. The impact that some parameters have on the predisposition to committing serial murders is shown using the data of actual persons in virtual computer experiments. Their results revealed some interesting regularities. For example, it was shown that if some maniac killers had children, this would reduce their predisposition to violence whereas for some others it would not. The same was observed when researching the impact of other parameters, such as the gravity of the psychiatric disorder, social and family status, etc. To determine how predisposed to serial crimes a certain person is, it is necessary to conduct a complex analysis of all the parameters in a mathematical model. Computer software based on this mathematical model was adapted for crime detection specialists. It is in open access on the website of Perm Branch of Artificial Intelligence Methodology Research Council of Russian Academy of Sciences. This software can be used by investigators of serial crimes at the initial stages of investigation when it is necessary to process large volumes of data regarding potential suspects.
Petri nets efficiently model both data- and control-flow. Control-flow is either modeled explicitly as flow of a specific kind of data, or implicit based on the data-flow. Explicit modeling of control-flow is useful for well-known and highly structured processes, but may make modeling of abstract features of models, or processes which are highly dynamic, overly complex. Declarative modeling, such as is supported by Declare and DCR graphs, focus on control-flow, but does not specify it explicitly; instead specifications come in the form of constraints on the order or appearance of tasks. In this paper we propose a combination of the two, using colored Petri nets instead of plain Petri nets to provide full data support. The combined approach makes it possible to add a focus on data to declarative languages, and to remove focus from the explicit control-flow from Petri nets for dynamic or abstract processes. In addition to enriching both procedural processes in the form of Petri nets and declarative processes, we also support a flow from modeling only abstract data- and control-flow of a model towards a more explicit control-flow model if so desired. We define our combined approach, and provide considerations necessary for enactment. Our approach has been implemented in CPN Tools 4.
Article considers theoretical prerequisites of creation of optimum hierarchical structure of system of monitoring of crucial parameters of food safety of Russia on the basis of application of the theory of indistinct sets.
This proceedings bring together contributions from researchers from academia and industry to report the latest cutting edge research made in the areas of Fuzzy Computing, Neuro Computing and hybrid Neuro-Fuzzy Computing in the paradigm of Soft Computing. The FANCCO 2015 conference explored new application areas, design novel hybrid algorithms for solving different real world application problems. After a rigorous review of the 68 submissions from all over the world, the referees panel selected 27 papers to be presented at the Conference. The accepted papers have a good, balanced mix of theory and applications. The techniques ranged from fuzzy neural networks, decision trees, spiking neural networks, self organizing feature map, support vector regression, adaptive neuro fuzzy inference system, extreme learning machine, fuzzy multi criteria decision making, machine learning, web usage mining, Takagi-Sugeno Inference system, extended Kalman filter, Goedel type logic, fuzzy formal concept analysis, biclustering etc. The applications ranged from social network analysis, twitter sentiment analysis, cross domain sentiment analysis, information security, education sector, e-learning, information management, climate studies, rainfall prediction, brain studies, bioinformatics, structural engineering, sewage water quality, movement of aerial vehicles, etc.
I give the explicit formula for the (set-theoretical) system of Resultants of m+1 homogeneous polynomials in n+1 variables