News reliability evaluation using Latent Semantic Analysis

Guo Xiaoning, Tan De Zhern, Soo Wooi King, Tan Yi Fei, Lam Hai Shuan

Research output: Contribution to journalArticleResearchpeer-review

2 Citations (Scopus)

Abstract

The rapid rise and widespread of 'Fake News' has severe implications in the society today. Much efforts have been directed towards the development of methods to verify news reliability on the Internet in recent years. In this paper, an automated news reliability evaluation system was proposed. The system utilizes term several Natural Language Processing (NLP) techniques such as Term Frequency-Inverse Document Frequency (TF-IDF), Phrase Detection and Cosine Similarity in tandem with Latent Semantic Analysis (LSA). A collection of 9203 labelled articles from both reliable and unreliable sources were collected. This dataset was then applied random test-train split to create the training dataset and testing dataset. The final results obtained shows 81.87% for precision and 86.95% for recall with the accuracy being 73.33%.

Original languageEnglish
Pages (from-to)1704-1711
Number of pages8
JournalTelkomnika
Volume16
Issue number4
DOIs
Publication statusPublished - 1 Aug 2018
Externally publishedYes

Keywords

  • Cosine similarity
  • Fake news detection
  • Latent semantic analysis
  • Natural language processing
  • Tf-idf

Cite this