Evaluating controversial articles on Wikipedia by talk page analysis
- Authors
- Avenda?o Sudario, Allan Roberto
- Format
- MasterThesis
- Status
- publishedVersion
- Description
This Master Thesis is aimed to provide an evaluation methodology for trustworthiness on collaborative systems. Currently, community managers or credible contributors require a sophisticated tools for supporting their decisions on breaking discussions. These tools should analyze in deep the underlying behavior of members reected through their uncon-scious acts when their interact among them. In particular, sentimental analysis is becom-ing one of the most trustworthy methodology for reecting the intention of the authors through their words used on their comments or conversation threads. The methodology shown here relies on text analysis through sentimental analysis techniques. In particular, the text studied here is extracted from Wikipedia' talk pages. For that, this thesis is divided in two main parts. The first part of this thesis covers the problem of identifying a certain point of agreement while authors interact on a discussion on Wikipedia talk pages. In the second part is considered the techniques for analyzing sets of controversial and featured pages. Here have been developed four approaches on different aspects to analyze: First, an analysis among controversial and featured topics based upon classifca-tion methods. Secondly, the analysis is refined through nine categories in common among controversial and featured topics. Third, it is proposed a methodology for determining whether a non-labeled page based upon models constructed on the first approach. The last approach is to identify some metrics by using Social Network Analysis.
- Publication Year
- 2014
- Language
- eng
- Topic
- SENTIMENT ANALYSIS
RANDOM WALK
WIKIPEDIA
COLLABORATIVE NETWORKS
- Repository
- Repositorio SENESCYT
- Rights
- openAccess
- License
- openAccess