Cookies Policy
We use cookies to improve our site and your experience. By continuing to browse our site you accept our cookie policy. Find out More
Close
  • Menu
About

About

Recieved my BSc in Mathematics, in 2013 and my MSc in Mathematics and Applications in 2015, both from University of Aveiro. Currently, I am a PhD student of Applied Mathematics (MAP-PDMA), working in the field of social network analysis.

Interest
Topics
Details

Details

  • Name

    Sofia Silva Fernandes
  • Cluster

    Computer Science
  • Role

    Research Assistant
  • Since

    01st September 2016
Publications

2018

Cover Image, Volume 8, Issue 5

Authors
Tabassum, S; Pereira, FSF; Silva Fernandes, Sd; Gama, J;

Publication
Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery

Abstract

2018

Dynamic graph summarization: a tensor decomposition approach

Authors
Fernandes, S; Fanaee T, H; Gama, J;

Publication
DATA MINING AND KNOWLEDGE DISCOVERY

Abstract
Due to the scale and complexity of todays' social networks, it becomes infeasible to mine them with traditional approaches. A possible solution to reduce such scale and complexity is to produce a compact (lossy) version of the network that represents its major properties. This task is known as graph summarization, which is the subject of this research. Our focus is on time-evolving graphs, a more complex scenario where the dynamics of the network also should be taken into account. We address this problem using tensor decomposition, which enables us to capture the multi-way structure of the time-evolving network. This property is unique and is impossible to obtain with other approaches such as matrix factorization. Experimental evaluation on five real world networks implies promising results demonstrating that tensor decomposition is quite useful for summarizing dynamic networks.

2018

Social network analysis: An overview

Authors
Tabassum, S; Pereira, FSF; Fernandes, S; Gama, J;

Publication
WILEY INTERDISCIPLINARY REVIEWS-DATA MINING AND KNOWLEDGE DISCOVERY

Abstract
Social network analysis (SNA) is a core pursuit of analyzing social networks today. In addition to the usual statistical techniques of data analysis, these networks are investigated using SNA measures. It helps in understanding the dependencies between social entities in the data, characterizing their behaviors and their effect on the network as a whole and over time. Therefore, this article attempts to provide a succinct overview of SNA in diverse topological networks (static, temporal, and evolving networks) and perspective (ego-networks). As one of the primary applicability of SNA is in networked data mining, we provide a brief overview of network mining models as well; by this, we present the readers with a concise guided tour from analysis to mining of networks. This article is categorized under: Application Areas > Science and Technology Technologies > Machine Learning Fundamental Concepts of Data and Knowledge > Human Centricity and User Interaction Commercial, Legal, and Ethical Issues > Social Considerations

2017

The Initialization and Parameter Setting Problem in Tensor Decomposition-Based Link Prediction

Authors
Silva Fernandes, Sd; Tork, HF; da Gama, JMP;

Publication
2017 IEEE International Conference on Data Science and Advanced Analytics, DSAA 2017, Tokyo, Japan, October 19-21, 2017

Abstract

2016

Stacked denoising autoencoders for the automatic recognition of microglial cells' state

Authors
Fernandes, S; Sousa, R; Socodato, R; Silva, L;

Publication
ESANN 2016 - 24th European Symposium on Artificial Neural Networks

Abstract
We present the first study for the automatic recognition of microglial cells' state using stacked denoising autoencoders. Microglia has a pivotal role as sentinel of neuronal diseases where its state (resting, transition or active) is indicative of what is occurring in the Central Nervous System. In this work we delve on different strategies to best learn a stacked denoising autoencoder for that purpose and show that the transition state is the most hard to recognize while an accuracy of approximately 64% is obtained with a dataset of 45 images.