A New Approach to Detect Communities in Multi-Weighted Co-authorship Networks Evelyn Perez Cervantes School of Systems Engineering National University of St Augustin, Peru epcervantes7@gmail.com Jes´ us P. Mena-Chalco Department of Computer Science University of S˜ ao Paulo, Brazil jmena@vision.ime.usp.br Abstract— Co-authorship graphs related to publications (bibliographical production) are generally represented by undi- rected graphs, where each edge maintains a single value for publications produced by the authors. A variation of this type of graph is related to use of different weights associated with different types of publications. In this context, within the field of scientometrics, ‘journal papers’ or ‘books’ may have a higher priority than the ‘conference papers’, and the ‘conference papers’ might have higher priority than the ‘extended abstract’. In this paper, we present a simple weight combination to detect communities in co-authorship networks, which allows simultaneous consideration of multiple types of collaborations. Our preliminary results show a good performance in the detection of communities considering real bibliographical production graphs. Keywords-Co-authorship network; community structure de- tection; multi-weighted graph I. I NTRODUCTION Recently, the algorithms for community structure de- tection have evolved and each time they use a greater amount of information. The first and most simple case is the communities detection in a binary and undirected co-authorship network. If two authors co-write a scientific paper, a unitary edge is created. Thus, the resulting graph is represented as an undirected unit-weighted graph. The second category is made up of binary and directed co- authorship networks, every edge in the undirected network is replaced by two, symmetrical and directed edge, the resulting graph is represented as a directed unit-weighted graph. This type of graph is commonly used to measure author prestige, but the binary graph does not represent co- authorships according to reality, so the most recent studies use weighted directed co-authorship networks, the weight of an edge is given by the number of co-authorship between the scientists who connect [1], [2], some of these studies in social networks do not use all information available in databases which leads to an unrealistic representation and detection of communities. In this work Newman’s algorithm [3] is used to detect community structure within real weighted networks, in order to make a more realistic representation and detection of communities was added the multi-weighted concept on the edges, which means that instead of representing the weight of an edge (the relationship between co-authors/scientists) Figure 1: Network with community structure. In this case there are three communities locked inside the large dashed circles, which have dense internal links but between which there are only a lower density of external links with a single value, it is represented by multiple values which have different weightings depending on the type of bibliographical production. This algorithm ensures keep up an special feature of social networks called community structure, the natural division of network nodes into groups within which the network connections are dense, with only a smaller number of edges between vertices of different groups as Figure 1. II. METHODOLOGY There are several proposals [4][5] to detect communities in graphs but we have not found any proposals that work with multiple weights in their edges or in this case with multiple types of bibliographical publication. For example, if there are two scientists who co-wrote 10 papers where 5 of them were published as ‘chapter books’, other 3 of them were published in ‘conference proceedings’ and the other 2 were published in another type of publication. It is natural to assign different priority levels to each group, e.g. the first group may have a weight of 3, and the next group a weight of 2 and the last group may have a weight of 1. The priority level is assigned arbitrarily depending on the relevance/importance of publications in