The paper concerns credibility of construction methods for affect-annotated lexicons, specifically a web questionnaire is explored and evaluated. Web-based surveys are susceptible to some risks, which might influence credibility of the results, as some participants might perform random clicks or intentionally falsify the responses. The paper explores the risks and proposes some strategies to reduce them. The strategies are supported by their experimental evaluation in the real case of SentiD affect-annotated dictionary construction.
Authors
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1109/acii.2015.7344605
- Category
- Aktywność konferencyjna
- Type
- materiały konferencyjne indeksowane w Web of Science
- Language
- angielski
- Publication year
- 2015
