Time-series topic analysis using singular spectrum transformation for detecting political business cycles

Abstract Herein, we present a novel topic variation detection method that combines extraction and change-point method. It extracts topics from time-series text data as the feature of each time detects change points changing patterns extracted topics. We applied this to analyze valuable, albeit underutilized, dataset containing Japanese Prime Minister’s (PM’s) detailed daily activities for over 32 years. The proposed provide insights into empirical analyses political business cycles, which is classical issue in economics science. For instance, our approach enables us directly observe PM’s actions, it can overcome challenges encountered by previous research owing unobservability behavior. Our observations are primarily consistent with recent theoretical developments regarding topic. Despite limitations, employing completely dataset, enhances understanding provides new classic issue.