14 Dec

Lengyel Balázs (MTA KRKT, Agglomeráció és Társadalmi Kapcsolathálózatok Lendület kutatócsoport vezető): Spatial diffusion and churn over the life-cycle of innovation: the role of social networks

Innovative ideas, products or services spread on social networks that, in the digital age, are maintained via telecommunication tools such as emails or social media. One of the last standing puzzles in social contagion is the role of physical space and it is not fully understood how products disappear from the map at the end of their life-cycle. In this paper, we utilize a unique dataset compiled from a Hungarian on-line social network (OSN) to uncover novel features in the spatial adoption and churn of digital technologies. The studied OSN was established in early 2000s and failed in international competition a decade later. A Bass Diffusion Model describes the process how the product gets adopted in the overall population. However, it does not cope with the prediction of spatial diffusion. The novel ingredients missing from the model are: the assortativity of adoption time, urban scaling of adoption over the product life-cycle and a distance decay function of diffusion probability. We find that early adopter towns also churn early; while individuals tend to follow the churn of nearby friends and are less influenced by the churn of distant contacts.

16 Nov

Gyimesi Péter (SZTE)

A szoftverhibák számos okból kifolyólag elkerülhetetlenek egy rendszer fejlesztése során: szűk határidők, pontatlan specifikáció, programozói figyelmetlenség, stb.

Ezeknek a hibáknak a megtalálása és kijavítása erőforrás igényes feladat. Számos kutatás foglalkozik a szoftverhibák felderítésével és különböző megközelítéseket alkalmaznak. Egy azonban közös bennük: a módszereket valahogyan tesztelni, azok eredményességét mérni kell. Ebből a célból közzétettek nyilvános hiba-adatbázisokat, melyek benchmark-ként szolgálnak az ilyen jellegű kutatások számára. Az előadás során az ilyen hiba-adatbázisok előállítására mutatunk egy hatékony módszert, mely egy gráf adatbázist (Neo4j) használ.

9 Nov

Ivan Luković: Formal Education in Data Science – A Perspective of Serbia and Faculty of Technical Sciences

In the last years, Data Science becomes an emerging education and research discipline all over the world. Software industry shows an increasing and even quite intensive interest for academic education in this area. The similar trend has been noticed in Serbia, particularly in Belgrade and Novi Sad. In this talk, we discuss main motivating factors for creating a new study program in data science at Faculty of Technical Sciences of University of Novi Sad. Also, we present a short survey of software industry needs for data science related experts, and discuss how we structured the new study program and addressed the main issues that come from more than evident industry requirements. The program was accredited in year 2015, both at the level of bachelor and master level studies, and this school year is its first execution, from which we expect the new experiences.

Vladimir Ivančević: An Overview of Selected Research Studies in Data Analytics and Data Science at the Faculty of Technical Sciences in Novi Sad, Serbia

Over the past several years, the Faculty of Technical Sciences in Novi Sad, Serbia, has been a home to an increasing number of research studies concerned with extraction of potentially valuable information from diverse data sets. Many of these research efforts were concentrated at the Chair of Applied Computer Science during a period in which data science started emerging as one of the most popular and promising new fields associated with computer science and informatics. The two most notable areas covered by these studies are education and medicine, both of which play an important role in the contemporary society and generate ample quantities of data. The education-focused studies dealt with a wide set of topics: a) exploring patterns in spatial distribution of students in a classroom, b) analysing student grades with respect to different factors, c) constructing programming tests automatically, and d) identifying position of engineering and technology education within the system of higher education in Serbia. The medicine-related studies centered upon epidemiology-oriented topics: a) creating a software system to support business intelligence in epidemiology and b) analysing collected data about early childhood caries to identify risk factors and create predictive models. Studies from both areas have provided some interesting findings and also managed to spur ideas for new research directions in theory and practice of data analytics and data science.

26 Oct

Sándor Szabó (University of Pécs): Estimating clique number in high performance computing environment

Suppose A is an algorithm to locate an upper estimate for the clique number of a given graph G. Using algorithm A one can construct a new A' algorithm that provides an improved estimate of the size of the maximum cliques in G. The fact that such algorithm A' exists does not come as a surprise to us. The point is that the construction is practical and the new algorithm A' well suited for various high performance computing environments. We carried out a large scale numerical experiment

to test our proposal.

28 Sept

Kristóf Kovács (BME): Facility location on networks, with hard to compute objective functions

This talk will be about facility location problems on networks with objective functions that are especially hard to compute. Calculating any point of these functions requires the solution of an NP-hard problem. Due to this property general global optimizer algorithms are inefficient to solve these problems. I will introduce the general Stackelberg problem, where two or more firms compete for demand by locating facilities one after the other. The choice of location for one of the firms influences the choice of the other firms, as both wants to maximize its profit after the facilities are built. Another hard to solve problem I will talk about is the 1-median problem with demand surplus, where one facility has to be located such that only a given percent of the demand has to be covered. Finally I will present the solution and the computational results to a Stackelberg problem, as well as a modified 1-median problem, both of which have hard to compute objective functions.