Detection and Time Series Variation of Latent Topic from Diary in Northern and Southern Courts Period of Japan

In the study, we detected the latent topics by Latent Dirichlet Allocation(LDA) and analyzed the texts through the analysis of the topics, with the text of “後愚昧言己” (Gogumai-ki) which is an old diary written in Northern and Southern courts period of Japan. We analyzed time-series variations of the topics and considered transitions of the topics in the text. In addition, we have tracked time series variations of the topics.