Infering trust in web-based social networks: an analysis from TidalTrust and T-SWEETS algorithms

Autores

  • Edeilson Milhomem
  • Diego Oliveira Rodrigues Universidade de Campinas (UNICAMP)
  • Ana Carolina Salgado Universidade Federal de Pernambuco (UFPE)
  • Silvio Romero Lemos Meira Universidade Federal de Pernambuco (UFPE)

DOI:

https://doi.org/10.20873/uft.2675-3588.2020v1n1p18

Palavras-chave:

trust inference, social network, recommender systems

Resumo

This work presents an analysis between two approaches for inferring trust in social networks: TidalTrust and T-SWEETS. The first part of this research is composed by an experiment with the Epinions, Slashdot and Wikipedia datasets in order to define the most suitable type of trust value (integer or fractioned, positive or negative, and so on). Furthermore, the second part of this research reports a comparative analysis between both algorithms. The results indicate that T-SWEETS outperforms TidalTrust in terms of accuracy and maintains the transitivity principle, which is the basic principle of trust.f trust.

Downloads

Publicado

2020-03-06

Como Citar

[1]
Milhomem, E. et al. 2020. Infering trust in web-based social networks: an analysis from TidalTrust and T-SWEETS algorithms . Academic Journal on Computing, Engineering and Applied Mathematics. 1, 1 (mar. 2020), 18–22. DOI:https://doi.org/10.20873/uft.2675-3588.2020v1n1p18.

Edição

Seção

Comunicação Breve

Categorias