A análise de redes sociais permite estudar a maneira como são estabelecidas as conexões entre indivíduos e como estas evoluem ao longo do tempo. A coautoria é uma das formas mais estudadas e bem documentadas de colaboração científica. Existem muitos aspectos de redes de colaboração científica, os quais podem ser rastreados de forma confiável através da análise de redes de colaboração usando métodos bibliométricos. Diversos esforços em diferentes áreas de pesquisa tentam analisar, entender, explicar e predizer o comportamento de sistemas modelados através de redes sociais. Nestes estudos, os indivíduos são modelados como vértices de um grafo, enquanto as relações entre eles são representadas por arestas. Atualmente, o estudo de redes de colaboração científica é importante e necessário para apoiar o planejamento estratégico, implementação e gestão dos programas de pesquisa científica. Neste trabalho, apresentamos um modelo de análise de redes científicas baseado em Grafos Relacionais com Atributos (ARG). O modelo proposto permite representar as redes de colaboração científica incluindo atributos individuais dos pesquisadores e atributos dos trabalhos colaborativos de pares de pesquisadores. Os dados correspondem às produções científicas de pesquisadores cadastrados na plataforma Lattes e extraídas automaticamente usando a ferramenta scriptLattes. Na primeira etapa, foi implementado o cálculo automatizado da taxa de internacionalização de cada pesquisador, a qual mostra a proporção entre o número de publicações internacionais e o número total de publicações. Esta medida junto com a produção científica individual discretizada em diversos grupos fazem parte das informações armazenadas nos vetores de atributos dos vértice dos ARGs. Por outro lado os vetores de atributos das arestas armazenam informações dos trabalhos colaborativos discretizados segundo a classificação da CAPES. Adicionalmente, neste trabalho foram exploradas duas aplicações relacionadas à (i) predição de trabalhos colaborativos futuros e à (ii) influência dos pesquisadores na rede de colaboração. O resultado da predição de vínculos foi usado para determinar a influência dos pesquisadores na redes de colaboração. A influência tem sido explorada com base na variação da predição de ligações com a presença ou a ausência do pesquisador na rede. Nossa proposta foi avaliada considerando diferentes testes sobre redes de coautoria científica de diversos grupos de pesquisadores. Os resultados obtidos são promissores para a análise de redes sociais em geral. / The social network analysis allows the study of how the relationships are established between individuals and how their are evolving with the time. The co-authorship is one of the most studied and documented scientific collaboration. There are some aspects which could be traced in a reliable way through the social network analysis using bibliometric methods. There are several proposals in different research areas trying to analyse, understand, explain and predict the behaviour of systems modeled as social networks. In this study, the individuals are modeled as vertices of a graph, while the relationships between them are represented by edges. Currently the study of scientific collaboration networks is important and necessary to support the strategic planning, implementation and management of scientific research programs. In this work, we present an scientific networks analysis model based on Attributed Relational Graphs (ARG). The proposed model allows to represent the scientific collaboration networks including individual attributes of researchers and attributes of the collaborative work of researchers pairs. The data correspond to the scientific production of researchers, registered in the Lattes Platform and automatically extracted using the tool scriptLattes \\citep{Mena-Chalco:2009}. In the first step, was implemented the automated computation of the internationalization rate for each researcher, that shows the ratio between the number of international publications and the total number of publications. This measure together with the individual scientific production discretized in diverse groups form part of the information stored in the vertices of the ARGs. On the other hand, the edges store information of collaborative work discretized according to the CAPES classification. Additionally, this work explores two related applications (i) prediction of future collaborative work and (ii) influence of researchers in collaboration network. The result of the link prediction was used to determine the influence of researchers in collaborative networks. The influence in collaboration network is computed based on the variation of the link prediction with the presence or absence of the researcher in the network. Our proposal was evaluated with different real scientific co-authorship networks and with different research groups. The results obtained look promising for analyzing social networks in general.

No description available.

The area of Knowledge Graphs has grown significantly during recent time and has found many different applications both in industrial and academic settings. Despite this, many large Knowledge Graphs are in fact incomplete, which leads to the problem of finding the missing facts in the graphs using Link Prediction. There are several ways of performing Link prediction, the most common one that has emerged recently being using Machine learning techniques to learn low-dimensional representations of the Knowledge Graph called Knowledge Graph embeddings. This project attempts to explore whether or not this is a viable method to use in order to give suggestions for companies that want to expand their businesses. In order to test this hypothesis, a Knowledge Graph was built using real company data from open sources. Then different Knowledge Graph embedding models were trained on the data in order to predict missing elements in the Knowledge Graph. The models were then compared to see which one is most suitable for this task and data set. The geometric based models were found to perform the best for the specific data set used in this project. In this category there are models such as TransE, TransR and RotatE. The results point to the method being a valid option for giving expansion suggestions to companies using a Knowledge Graph of other companies and their products. However, to be certain of this, further research needs to be done where the method needs to be implemented on a larger scale using more diverse data. / Området kunskapsgrafer har växt mycket under de senaste åren och har många olika tillämpningar både inom akademiska och industriella områden. Trots denna tillväxt så är många kunskapsgrafer ofullständiga, vilket leder till problemet att hitta den faktan i kunskapsgraferna som saknas genom något som kallas länkförutsägelser. Det finns många olika metoder för att göra länkförutägelser, men den populäraste metoden som uppkommit de senaste åren är att använda maskininlärning för att lära in lågdimensionerade representationer av kunskapsgrafen i något som kallas kunskapsgrafsinbäddningar. I det här projektet försöker vi ta reda på om den här metoden går att använda för att ge förslag för företag som vill expandera och etablera sig på nya marknader. För att testa om detta är möjligt byggdes en kunskapsgraf med hjälp av data från öppna källor. Sedan fick olika kunskapsgrafsinbäddningsmodeller träna på data från kunskapsgrafen för att sedan kunna hitta fakta i grafen som saknades. De olika modellerna jämfördes sedan för att se vilken som var mest lämplig för att klara av uppgiften på vår kunskapsgraf. De modeller som är geometribaserade visade sig prestera bäst, bland dom fanns modeller som TransE, TransR och RotatE. Resultaten från projektet visar på att metoden är användbar för uppgiften att ge förslag om områden som ett företag kan expandera till. Dock skulle detta behöva undersökas mer med en större mer mångfaldig mängd data för att vara säker på att detta går att använda i fler marknadsområden än dem som ingick i projektet.

No description available.