Sarka Zehnalova

Local representatives in weighted networks

The main features of current real-world networks are their large sizes and structures, which show varying degrees of importance of the nodes in their surroundings. The topic of evaluating the importance of the nodes offers many different approaches that usually work with unweighted networks. We present a novel, simple and straightforward approach for the evaluation of the networks nodes with a focus on local properties in their surroundings. The presented approach is intended for weighted networks where the weight can be interpreted as the proximity between the nodes. Our suggested x-representativeness then takes into account the degree of the node, its nearest neighbors and one other parameter which we call the x-representativeness base. Following that, we also present experiments with three different real-world networks. The aim of these experiments is to show that the x-representativeness can be used to deterministically reduce the network to differently sized samples of representatives, while maintaining the topological properties of the original network.

Local Dependency in Networks

Abstract Many real world data and processes have a network structure and can usefully be represented as graphs. Network analysis focuses on the relations among the nodes exploring the properties of each network. We introduce a method for measuring the strength of the relationship between two nodes of a network and for their ranking. This method is applicable to all kinds of networks, including directed and weighted networks. The approach extracts dependency relations among the network’s nodes from the structure in local surroundings of individual nodes. For the tasks we deal with in this article, the key technical parameter is locality. Since only the surroundings of the examined nodes are used in computations, there is no need to analyze the entire network. This allows the application of our approach in the area of large-scale networks. We present several experiments using small networks as well as large-scale artificial and real world networks. The results of the experiments show high effectiveness due to the locality of our approach and also high quality node ranking comparable to PageRank.

Email Conversation Network Analysis: Work Groups and Teams in Organizations

Email communication is a source of important information, much of which is at first sight hidden. This paper presents an analytical tool that was created to analyze the deeper relationships in the email data. Those include relationships based on an interaction of multiple users in a team. The analytical methods proposed and described in this paper are based on two factors. The first factor is the context, which is a group of multiple users in combination with terms used in the communication. The second factor is the time interval in which the communication was conducted. Based on these factors, we analyze the conversations that take place and get results that are in several different forms presented to the users. The paper presents methods for weighting conversations, users and relationships, as well as methods for finding communities associated with the specified context. Additionally, the concept of the explorative user interface is introduced.

Evolution of Author's Topic in Authorship Network

There may be several reasons why people publish together. Above all, the fact that the authors share common professional interests is the main reason. In our research we work with the DBLP dataset which contains the basic bibliographic information of publications from the computer science field. These data are freely available and contain highly relevant information about publication activity from the period of nearly fifty years, even though they are not complete. One of the goals of our research is to analyze and visualize the evolution of authors and co-authorship from the point of view of research topics. We present the results of our research in this paper. One of the results is also visualization in our online FORCOA.NET system.

Quantitative assessment of informative immunophenotypic markers increases the diagnostic value of immunophenotyping in mature CD5-positive B-cell neoplasms: QUANTITATIVE IMMUNOPHENOTYPING OF CD5+ B-CELL NEOPLASMS

The data on the clinical utility of the quantitative assessment of immunophenotypes in distinguishing mature CD5‐positive B‐cell neoplasms is limited. The study aim was to assess the diagnostic value of the quantitative assessment of a panel of 18 markers and to identify the most informative ones.

Graph Construction Based on Local Representativeness.

Graph construction is a known method of transferring the problem of classic vector data mining to network analysis. The advantage of networks is that the data are extended by links between certain (similar) pairs of data objects, so relationships in the data can then be visualized in a natural way. In this area, there are many algorithms, often with significantly different results. A common problem for all algorithms is to find relationships in data so as to preserve the characteristics related to the internal structure of the data. We present a method of graph construction based on a network reduction algorithm, which is found on analysis of the representativeness of the nodes of the network. It was verified experimentally that this algorithm preserves structural characteristics of the network during the reduction. This approach serves as the basis for our method which does not require any default parameters. In our experiments, we show the comparison of our graph construction method with one well-known method based on the most commonly used approach.

Local representativeness in vector data.

The amount of large-scale real data around us is increasing in size very quickly, as is the necessity to reduce its size by obtaining a representative sample. Such sample allows us to use a great variety of analytical methods, the direct application of which on original data would be unfeasible. Conventional sampling methods provide non-deterministic results trying to preserve selected characteristics of the input dataset. We present a novel, simple, straightforward and deterministic approach with the same goal. It is not sampling in the true sense but a reduction of vector data, which maintains very well internal data structures (clusters and density). The approach is based on analyzing the nearest neighbors. Our suggested x-representativeness then takes into account the local density of the data and nearest neighbors of individual data objects. Following that, we also present experiments with two different datasets. The aim of these experiments is to show that the x-representativeness can be used to deterministically reduce the datasets to differently sized samples of representatives, while maintaining properties of the original datasets.

Using self-organizing maps for identification of roles in social networks

In social networks the participants may be characterized by their roles. We understand roles as different patterns of link structure in the network. These roles describe the node and its activity in the network over time. Self-organizing maps (SOMs) - type of artificial neural-networks, are used for nodes role identification and for discovery of all the roles present in the network. Different data preprocessing methods allow us to capture different aspects of roles. We show results of the experiment with a large scale co-authorship network constructed from a DBLP dataset.

Comparing two local methods for community detection in social networks

One of the most obvious features of social networks is their community structure. Several types of methods were developed for discovering communities in the networks, either from the global perspective or based on local information only. Local methods are appropriate when working with large and dynamic networks or when real-time results are expected. In this paper we explore two such methods and compare the results obtained on the sample of a co-authorship network. We study how much may detected communities vary according to the method used for computation.

Dynamic changes in HLA-DR expression during short-term and long-term ibrutinib treatment in patients with chronic lymphocytic leukemia

There is the first evidence of changes in the kinetics of B cell antigen receptor (BCR) internalisation of neoplastic cells in chronic lymphocytic leukemia (CLL) after the short-term and long-term administration of ibrutinib. We aimed to assess the influence of short-term and long-term ibrutinib treatment on the HLA-DR expression on CLL cells, T cells and monocytes. The immunophenotyping of CLL and immune cells in peripheral blood was performed on 16 high-risk CLL patients treated with ibrutinib. After early ibrutinib administration, the HLA-DR expression on CLL cells reduced (P = 0.032), accompanied by an increase in CLL cell counts in peripheral blood (P = 0.001). In vitro culturing of CLL cells with ibrutinib also revealed the reduction in the HLA-DR expression at protein and mRNA levels (P < 0.01). The decrease in HLA-DR on CLL cells after the first month was followed by the gradual increase of its expression by the 12th month (P = 0.001). A one-month follow-up resulted in elevated absolute counts of CD4+ (P = 0.002) and CD8+ (P < 0.001) T cells as well as CD4+ and CD8+ cells bearing HLA-DR (P < 0.01). The long-term administration of ibrutinib was associated with the increased numbers of CD4+ bearing HLA-DR (P = 0.006) and elevation of HLA-DR expression on all monocyte subsets (P ≤ 0.004). Our results provide the first evidence of the time-dependent immunomodulatory effect of ibrutinib on CLL and T cells and monocytes. The clinical consequences of time-dependent changes in HLA-DR expression in ibrutinib treated patients deserve further investigation.