An ANN-based approach of interpreting user-generated comments from social media

T. C. Wong, Hing Kai Chan, Ewelina Lacka

Research output: Contribution to journalArticle

5 Citations (Scopus)
29 Downloads (Pure)

Abstract

The IT advancement facilitates growth of social media networks, which allow consumers to exchange information online. As a result, a vast amount of user-generated data is freely available via Internet. These data, in the raw format, are qualitative, unstructured and highly subjective thus they do not generate any direct value for the business. Given this potentially useful database it is beneficial to unlock knowledge it contains. This however is a challenge, which this study aims to address. This paper proposes an ANN-based approach to analyse user-generated comments from social media. The first mechanism of the approach is to map comments against predefined product attributes. The second mechanism is to generate input-output models which are used to statistically address the significant relationship between attributes and comment length. The last mechanism employs Artificial Neural Networks to formulate such a relationship, and determine the constitution of rich comments. The application of proposed approach is demonstrated with a case study, which reveals the effectiveness of the proposed approach for assessing product performance. Recommendations are provided and direction for future studies in social media data mining is marked.
Original languageEnglish
Pages (from-to)1169-1180
Number of pages12
JournalApplied Soft Computing
Volume52
Early online date23 Sep 2016
DOIs
Publication statusPublished - 31 Mar 2017

Keywords

  • neural network application
  • statistical methods
  • product performance
  • social media
  • information exchange
  • user-generated data
  • ANN-based approach
  • predefined product attributes
  • user-generated comments
  • input-output models
  • comment length
  • data mining

Fingerprint Dive into the research topics of 'An ANN-based approach of interpreting user-generated comments from social media'. Together they form a unique fingerprint.

Cite this