The use of text mining for obtaining new knowledge from historical documents has gained wide attention in humanities research. We propose a method of revealing and visualizing relationships between historical persons using locational information expressed in historical documents. For each person, a vector of co-occurring place names is constructed. We then clustered persons using this vector in the K-means algorithm. We conducted experiments using
, a diary written by an aristocrat in the classical era. The result showed that the obtained clusters match well with historically meaningful groups, indicating the effectiveness of our proposed method.
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten