This, the second issue of Volume 12 of the Computer Science and Information Systems journal, consists of nine regular articles and two special sections: Special Section on Advances in Systems, Modeling and Languages, and Special Section on Advances in Information Technology. We thank all guest editors, authors and reviewers for the hard work and enthusiasm which were invested into preparing the current issue of our journal.

We are happy to announce the new two-year impact factor of our journal for 2014, which is now 0.477. In addition, ComSIS has also been assigned a five-year impact of 0.547.

The first regular article, “A Lane-Changing Behavioral Preferences Learning Agent with its Applications” by Wang Jian et al., studies lane-changing (LC) behavior of drivers, proposing an extension to the traditional belief-desire-intention (BDI) agent model which can perform behavioral preference learning in the LC process – the behavioral preferences learning-BDI (DpL-BDI) model. The introduced behavioral preference learning strategies gradually grasp the essentials in driver’s subjective judgments in LC decision-making, while colored Petri nets are used to realize the components and scheduler of DpL-BDI agents.

Sohaib Hamioud and Fadila Atil, in “Model-Driven Java Code Refactoring,” introduce a model-driven approach to code refactoring where refactoring features (code representation, analysis and transformation) adopt models as first-class artifacts, with the aim of exploring the value of model transformation and code generation when formalizing refactorings and developing tool support.

“Edge-texture 2D image quality metrics suitable for evaluation of image interpolation algorithms” by Sanja Maksimović-Moićević et al. proposes new objective full-reference metrics of image quality, with the aim of providing a better match with subjective image quality assessments. The method consists two quality measures which separately indicate image quality on edges and in texture areas, and is particularly suited for evaluation of interpolation and compression algorithms.

In their paper entitled “Optimization Design of Traffic Flow under Security Based on Cellular Automata Model,” Fan Zhang et al. propose a cellular automata model to optimize traffic flow as to maximize the flux on each road whilst ensuring safety. Different settings are investigated: three and four-lane highways, different velocities and densities, rear-end and lateral collisions, reaction times and distances, and different vehicles. Experimental results indicate that the number of lanes is not an influential factor, while the maximum speed limit plays a significant role in light traffic, and the minimum speed limit is of great importance in heavy traffic.

“Adaptive Mobile Visualization – The Chameleon Framework,” by Paulo Pombinho et al., tackles the problem of information visualization on mobile devices, where limited screen space, resource constraints and interaction restrictions pose difficulties for both developers and users. Addressing these problems by adapting the visualization to the user context has so far been done through ad-hoc approaches that are difficult to generalize. This paper presents a framework for adaptive mobile visualization (AMV) applications, named Chameleon, and the development and evaluation of prototypes that use it.

Dejan Mančev and Branimir Todorović, in “k-Best Max-margin Approaches for Sequence Labeling,” introduce four new k-best extensions of max-margin structured algorithms for sequence labeling, discuss their properties and connection, and evaluate all algorithms on two problems – shallow parsing and named entity recognition, showing that the algorithms are affected by the changes of the k parameter, and offer improvements compared to the single best case.

Jovana Božić and Đorđe Babić, in their paper “EUR/RSD Exchange Rate Forecasting Using Hybrid Wavelet-Neural Model: A CASE STUDY,” examine and discuss modeling and prediction results of several exchange rates, with the main focus on EUR/RSD, using a combination of wavelet transforms, neural networks and statistical time series analytical techniques. Although two resolution levels are used to decompose financial time series into a wavelet representation, results show that increase in resolution does not improve prediction accuracy. The results also indicate that the model sufficiently satisfies the characteristics of a financial predictor.

In “Business Process Model and Notation: The Current State of Affairs,” Mateja Kocbek et al. present a systematic literature overview in order to provide the state-of-the-art results addressing the acceptance of business process model and notation (BPMN), and the purposes of its use. The findings suggest that BPMN is still widely perceived as the de facto standard in the process modeling domain and its usage is increasing, but that only a limited set of elements are commonly used, with several extensions being proposed.

And finally, Carlos Blanco et al., in “An MDA approach for developing Secure OLAP applications: metamodels and transformations,” extend their previously proposed model-driven approach for developing a secure data warehouse repository by improving UML profiles for conceptual modeling, defining a logical metamodel for OLAP applications, and defining and implementing transformations from conceptual to logical models.

Mirjana Ivanović

Managing Editor
Miloš Radovanović