Accommodation reviews are valuable resources for
future guests to know the opinions of users who have already
stayed at a particular place before. However, it is difficult for
users to extract the information specific to each topic such as
facilities, access, and breakfast. We consider that the seasonal
features of accommodations are especially important to ensure a
comfortable and enjoyable stay. This paper proposes a hierarchical topic analysis with time variation to extract seasonal features
for accommodations. The proposed method extracts seasonally
important words and shows the similarity of topics that the
important words belong to between seasons. In this paper, we
discuss the effectiveness of the extracted features as references
for guests to choose accommodations.