A) The Proceedings will be published in Digital Form.

B) Extended versions of the accepted Papers will be evaluated for possible publication in the following Journals
indexed in many reputable indexes, like Web of Science (Database: Emerging Sources Citation Index), SCOPUS, EI Compendex, IET, Google Scholar.
For the indexing of a particular journal visit its web page.

International Journal on Advanced Science, Engineering and Information Technology

http://ijaseit.insightsociety.org/

ARPN Journal of Engineering and Applied Sciences

http://www.arpnjournals.com/jeas/

International Journal of Applied Mathematics and Computer Science

https://www.amcs.uz.zgora.pl/

International Journal of Computers and Applications (Taylor and Francis)

http://www.tandfonline.com/toc/tjca20/current

Journal of Electrical Systems (JES)

http://journal.esrgroups.org/jes/

Journal of Theoretical and Applied Information Technology

http://www.jatit.org

Advances in Electrical and Electronic Engineering

http://advances.utc.sk/index.php/AEEE/index

TELKOMNIKA: Telecommunication, Computing, Electronics and Control

http://telkomnika.ee.uad.ac.id/

International Journal of Modern Manufacturing Technologies

http://www.ijmmt.ro/subscriptionsandexchanges.php

Indonesian Journal of Electrical Engineering and Computer Science (IJEECS)

http://www.iaescore.com/journals/index.php/IJEECS/index

Bulletin of Electrical Engineering and Informatics (BEEI)

http://journal.portalgaruda.org/index.php/EEI/index

Proceedings of the Institute of Mathematics and Mechanics

http://proc.imm.az/

Journal of Fundamental and Applied Sciences (JFAS)

http://www.jfas.info/index.php/jfas

Journal of Electrical Systems (JES)

http://journal.esrgroups.org/jes/

International Journal of Engineering and Technology

https://www.sciencepubco.com/index.php/ijet/index

International Journal of Power and Energy Systems

http://www.actapress.com/Content_of_Journal.aspx?JournalID=233

International Journal of Robotics and Automation

http://www.actapress.com/Content_of_Journal.aspx?JournalID=237

Reports of Biochemistry and Molecular Biology

http://rbmb.net/

Remote Sensing in Ecology and Conservation

https://zslpublications.onlinelibrary.wiley.com/journal/20563485

Quarterly Journal of Finance

https://www.worldscientific.com/worldscinet/qjf

International Journal of Electrical and Computer Engineering (IJECE)

http://iaesjournal.com/online/index.php/IJECE

International Journal of Power Electronics and Drive Systems

http://iaesjournal.com/online/index.php/IJPEDS

or in other similar journals indexed in:
Web of Science (Database: Emerging Sources Citation Index), SCOPUS, EI Compendex, IET, Google Scholar.

**
* Contact us by email:
support@inase.org ****Registration fees
Publication Ethics and Malpractice Statement
Deadlines**

(Previous Conference: PMAMCM 2017, Agia Pelagia Beach, Crete, Greece, July 14-17, 2017)

### PLENARY SPEAKERS:

**Prof. Dana Simian, University "Lucian Blaga" of Sibiu, Faculty of Sciences, Research Center on Informatics and Information Technology, ROMANIA, e-mail: dana.simian@ulbsibiu.ro**

**Title:**"Optimization of Multiple Kernels in Support Vector Machines for Classification and Regression (SVMs and SVR)"

**Abstract:**The aim of this talk is to present a general approach for building optimal multiple kernels and for automatic optimization of the parameters in classification and regression based on support vectors, ie in Support Vector Machines (SVMs) and Support Vector Regression (SVR). SVMs are supervised learning methods introduced by Vapnik. Initialy SVMs were design for solving binary classification for linear separable data. SVMs construct, based on a training set of data, a linear classifyer with maximum margin. Training examples that are closest to this maximal separating hyperplane are named support vectors and they represents the only examples (vectors) which contribute to the expression of the classification hyperplane. SVMs were adapted for multi-class problems. For solving non-linear classification problems the data are mapped in a higher dimensional space where they become linearly separable. Using the "kernel trick" it is not necessary effectively to know this mapping. A kernel function is used for expressing the inner products involved in the equation of the separating hyperplane. SVR is a version of a SVM for regression. Both in SVM and SVR the choice of the kernel is extremely important for the prediction performances. Standard SVMs use one single kernel and the prediction supposes the choice of the kernel parameters. Usually the choice of the kernel is made in an empirical way. The real problems require more complex kernels. We propose and present an approach for building optimal multiple SVM and SVR kernels and automatic select optimal model parameters. We formalize our method and provide real applications and the conclusions of a sensitivity study.

**Prof. Janusz Brzdek, Department of Mathematics, Pedagogical University of Cracow, POLAND, e-mail: jbrzdek@up.krakow.pl**

**Title:**"Approximate Solutions of Equations and Ulam Stability"

**Abstract:**The following natural question arises in many areas of scientific investigations: what errors we commit replacing the exact solutions to some equations by functions that satisfy those equations only approximately (or vice versa). Some efficient tools to evaluate those errors can be found in the theory of Ulam stability. The issue of Ulam stability of an equation can be very roughly expressed in the following way: when must a function satisfying an equation approximately (in some sense) be close (in a given way) to an exact solution to the equation? The problem of such stability was formulated for the first time by Ulam in 1940 for homomorphisms of metric groups; a solution to it was published a year later by Hyers for Banach spaces. Again in 1960, Ulam asked the following general question: when is it true that a solution of an equation differing slightly from a given one, must of necessity be close to a solution of the given equation? We discuss several issues connected to those questions for various (difference, differential, integral, functional) quations.

**Prof. Ljiljana Petrovic, Department of Mathematics and Statistics, Faculty of Economics, University of Belgrade, SERBIA, e-mail: petrovl@ekof.bg.ac.rs**

**Title:**"Some Generalization of Granger Causality and Stochastic Dynamic Systems"

**Abstract:**The basic idea in this paper is to relate some concepts of causality to the stochastic realization problem. More precisely, we consider the next problem (that follows directly from realization problem): how to find Markovian representations (even minimal) for a given family of Hilbert spaces (understood as outputs of a stochastic dynamic system S1 provided it is in a certain causality relationship with another family of Hilbert spaces (i.,e. with some informations about states of a stochastic dynamic system S2. Granger causality (C.W.J. Granger, Investigation Causal Relations by Econometric Models and Cross Spectral Methods, Econometrica. 37, 1969, 424-438) is one of the most popular measures to reveal causality influence of time series widely applied in economics, demography, neuroscience etc. The study of Granger-causality has been mainly preoccupied with time series. We shall instead concentrate on continuous time processes because many of systems to which it is natural to apply tests of causality, take place in continuous time. For example, this is generally the case within economy, finance, physics, medicine etc. First, we give various concepts of causality relationship between flows of information represented by families of Hilbert spaces. We then relate some concepts of causality to the stochastic realization problem. The approach adopted in this paper is that of [3]. However, since our results do not depend on probability distribution, we deal with arbitrary Hilbert spaces instead of those generated by Gaussian processes. Let us suppose that a stochastic dynamic system S1 causes, in a certain sense, behavior of some other stochastic dynamic system S2 . It is natural to assume that outputs H of system S1 can be registered and that some information E about states (or perhaps states themselves) of system S2 is given. Results we shall prove will tell us under which conditions concerning the relationships between H and E it is possible to find states\ G of system S1 which are in a certain causality relationship with H and E. It is clear that all given results can be extended to the case when a flows of informations are families of σ-algebras generated by finite dimensional Gaussian random variables. But, in the case that σ-algebras are arbitrary, not necessarily generated by Gaussian random variables the extensions of the proofs from this paper is nontrivial because one can not take an orthogonal complement with respect to a σ-algebra as one can with respect to subspace in Hilbert space.