We are especially grateful to Kaito Yamauchi, who was a master student of the first author, for his contribution to the initial idea of using a dynamic topic model to classify texts and extract topics from news articles. He graduated from Kwansei Gakuin University in 2013 with a master’s thesis entitled “Stock market analysis using a time series topic model” in Japanese.

This study is partly supported by the Institute of Statistical Mathematics (ISM) cooperative research program (2013-ISM-CRP-4202, 2014-ISM-CRP-4102, 2015-ISM-CRP-2012 and 2016-ISM-CRP-2011) and JSPS KAKENHI Grant Number 26380279 and 15K03406.