Historical inference based on semi-supervised learning
作者：
Highlights：
• We propose a framework of classifying people in history into rivalry power groups (parties).
• The proposed method employs graph-based semi-supervised learning.
• To create a network from genealogy, we propose a method for converting the tree structure to a network.
• We devise a labeling method using historical records on political decisions.
• The paper is a pioneering work of machine learning applied to history, which can help people infer the unrevealed facts in history.
摘要
•We propose a framework of classifying people in history into rivalry power groups (parties).•The proposed method employs graph-based semi-supervised learning.•To create a network from genealogy, we propose a method for converting the tree structure to a network.•We devise a labeling method using historical records on political decisions.•The paper is a pioneering work of machine learning applied to history, which can help people infer the unrevealed facts in history.
论文关键词：Machine learning,Semi-supervised learning,Historical big data,Genealogy
论文评审过程：Received 20 September 2017, Revised 14 December 2017, Accepted 28 March 2018, Available online 5 April 2018, Version of Record 13 April 2018.
论文官网地址：https://doi.org/10.1016/j.eswa.2018.03.059
