András London

Complex network analysis of public transportation networks: A comprehensive study

In this study, using the network approach, we analyzed the urban public transportation systems of 5 Hungarian cities. We performed a comprehensive network analysis of the systems with the main goal of identifying significant similarities and differences of the transportation networks of these cities. Although previous studies often investigated unweighted networks, one novelty of our study is to consider directed and weighted links, where the weights represent the capacities of the vehicles (bus, tram, trolleybus) in the morning peak hours. In particular, we calculated descriptors of global network characteristic and various centrality measures of the network nodes in both the weighted case and unweighted case. By comparing the results obtained for the different cities, we get a highly detailed picture of the differences in the organization of the public transport, which may due to historical and geographical factors. Also, by comparing the results obtained from the weighted and unweighted approaches, we can identify which are the most sensitive routes and stations of the network pointing out some organizational inconsistencies of the transportation system.

Time-dependent network algorithm for ranking in sports

In this paper a novel ranking method which may be useful in sports like tennis, table tennis or American football, etc. is introduced and analyzed. In order to rank the players or teams, a time-dependent PageRank based method is applied on the directed and weighted graph representing game results in a sport competition. The method was examined on the results of the table tennis competition of enthusiastic sport-loving researchers of the Institute of Informatics at the University of Szeged. The results of our method were compared by several popular ranking techniques. We observed that our approach works well in general and it has a good predictive power.

Student evaluation by graph based data mining of administrational systems of education

In this paper, we define a modified PageRank algorithm as a data mining technique with the aim of evaluating the achievements of students and generate a ranking between them. We applied our method to the data set of a complex administrational system that contains numerous well detailed data of several schools in public education. In order that the method can be applied, we constructed a directed, weighted network of students, where edges of the network represent the comparability of the students in an appropriate way. We compared the results of our method with the standard statistical techniques that are used for rating and ranking the students and observed that our method gives a clearer picture about their educational achievements. Further advantages of graph based data mining techniques in educational systems are also highlighted.

HITS based network algorithm for evaluating the professional skills of wine tasters

Two popular and widely used webpage ranking algorithms are PageRank and HITS. We considered the 2009 Szeged Wine Fest data and another reliable data set of wines from the famous region Villány and, on basis of each data set, constructed a directed and weighted bipartite graph of wine tasters and wines. We applied an extended version of PageRank and HITS, the Co-HITS algorithm to wine tasting graph in order to rank tasters according to their ability and professional skill. The results of our technique were compared to other simple statistical methods. In general we observed that our ranking method performed better: it can filter out incompetent tasters, who, for example, gave the average score of some other tasters for the wines she or he tasted. Furthermore, our method gives a clearer picture about the competence of wine tasters.

Core of communities in bipartite networks

We use the information present in a bipartite network to detect cores of communities of each set of the bipartite system. Cores of communities are found by investigating statistically validated projected networks obtained using information present in the bipartite network. Cores of communities are highly informative and robust with respect to the presence of errors or missing entries in the bipartite network. We assess the statistical robustness of cores by investigating an artificial benchmark network, the coauthorship network, and the actor-movie network. The accuracy and precision of the partition obtained with respect to the reference partition are measured in terms of the adjusted Rand index and the adjusted Wallace index, respectively. The detection of cores is highly precise, although the accuracy of the methodology can be limited in some cases.

Applying graph-based data mining concepts to the educational sphere

In this study, we discuss the possible application of the ubiquitous complex network approach for information extraction from educational data. Since a huge amount of data (which is detailed as well) is produced by the complex administration systems of educational institutes, instead of the classical statistical methods, new types of data processing techniques are required to handle it. We define several suitable network representations of students, teachers and subjects in public education and present some possible ways of how graph mining techniques can be used to get detailed information about them. Depending on the construction of the underlying graph, we examine several network models and discuss which are the most appropriate graph mining tools (like community detection and ranking and centrality measures) that can be applied on them. Lastly, we attempt to highlight the many advantages of using graph-based data mining in educational data against the classical evaluation techniques.

Spanning Tree Game as Prim Would Have Played

In this paper, we investigate special types of Maker-Breaker games defined on graphs. We restrict Maker’s possible moves that resembles the way that was introduced by Espig, Frieze, Krivelevich and Pedgen [9]. Here, we require that the subgraph induced by Maker’s edges must be connected throughout the game. Besides the normal play, we examine the biased and accelerated versions of these games.

The structure of pairing strategies for k-in-a-row type games

In Maker-Breaker positional games two players, Maker and Breaker, play on a finite or infinite board with the goal of claiming or preventing the opponent from getting a finite winning set, respectively. For different games there are several winning strategies for Maker or Breaker. One class of winning strategies is the so-called pairing (paving) strategies. Here, we describe all possible pairing strategies for the 9-in-a-row game. Furthermore, we define a graph of the pairings, containing 194,543 vertices and 532,107 edges, in order to give them a structure. A complete characterization of the graph is also given.

A világkereskedelem hálózatelméleti vizsgálatának lehetőségeiről

Tanulmanyunkban a vilagkereskedelmi rendszer vizsgalatanak egy, az utobbi időben egyre nepszerűbb halozatelemzesen alapulo, elsősorban grafelmeleti modszereket alkalmazo megkozeliteset mutatjuk be a szakirodalom es szemleletes peldak alapjan. Olyan komplex halozatos modelleket kivanunk ismertetni, amelyekkel vizsgalni lehet a nemzetkozi kereskedelmi rendszer - mint komplex halozat - belső tulajdonsagait es fejlődeset. Attekintjuk a rendszer lehetseges grafos reprezentacioit, ismertetunk nehany halozatelemzesi modszert, amelyek segitsegevel vazoljuk az informaciokinyeres lehetősegeit. Journal of Economic Literature (JEL) kod: C02, C38, C53.

A new model for the linear 1-dimensional online clustering problem

Abstract In this study, a mathematical model is presented for an online data clustering problem. Data clustering plays an important role in many applications like handling the data acknowledgment problem and data stream management in real-time locating systems. The inputs in these problems are data sequences, each containing several data elements. Each data element has an arrival time and a weight that reflects its importance. The arrival times are not known in advance, and some data elements never arrive. Hence the system should decide which moment is optimal for forwarding the collected data for processing. This requires finding a good trade-off between the amount of collected information and the waiting time, which may be regarded as a minimization problem. Here, we investigate several online algorithms and present their competitive analysis and average case studies. Experimental results, based on simulations using artificially generated data, are also presented and they confirm the efficiency of our methods.