How to find people who can help to answer a question? – Analyses of metrics and machine learning in online communities
作者：
Highlights：
• The Core component has similar results to the whole network.
• The number of answers is the best metric to discover the user reputation.
• The users’ attributes in the Core component is a better approach than graph metrics.
• A previously trained neural network may be applied to other community.
• Around 90% of the users were correctly identified using neural networks.
摘要
•The Core component has similar results to the whole network.•The number of answers is the best metric to discover the user reputation.•The users’ attributes in the Core component is a better approach than graph metrics.•A previously trained neural network may be applied to other community.•Around 90% of the users were correctly identified using neural networks.
论文关键词：Collaborative learning,Online communities,Experts,Reputation,Bow Tie structure,Machine learning
论文评审过程：Available online 29 December 2014, Version of Record 29 July 2015.
论文官网地址：https://doi.org/10.1016/j.chb.2014.12.026
