The method of trust level of publications hosted in virtual communities

Abstract

The proposed model of data collection and analysis from thematic virtual communities using known information analysis techniques: scoring and parsing. Open communities were selected for the study, namely their architecture and main components: information content (title, description, posts, topics of the event) and audience (community members). To select relevant, informative, reliable publications, the scoring method is used which reflects the level of trust of the authors of the publication in the form of weighted indicators of a set of certain characteristics. Data collection is a combined approach, as virtual communities are dynamic in the content of the data and their content depends on the actions of the participants. To parse posts from virtual communities, it was decided to use ImportXML function in Microsoft Excel, which allows you to collect data from different sources, and then sample, analyze, and select the presentation of results using other built-in tools of this program.