Berlin 2018 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

SOE: Fachverband Physik sozio-ökonomischer Systeme

SOE 22: Focus Session: Computational Social Science

SOE 22.4: Vortrag

Donnerstag, 15. März 2018, 17:00–17:15, MA 001

Two types of seasonal words observed from Japanese blog data — •Kenta Yamada — National Institute of Informatics, Tokyo, Japan — Precursory Research for Embryonic Science and Technology Japan Science and Technology Agency Saitama, Japan

I analyzed frequency of word appearances in Japanese blogs and introduced the method which detects two types of seasonal words using simple autocorrelation analysis: one is for seasonal words with a specific day such as Christmas basically having sharp growth and decay around the peaking day characterized by a power function, and the other one is for seasonal words without a specific day like ski. The algorithm caught not only words which are easily understood as seasonal words such as Christmas and ski but also words which are not well known by everyone such as words related to local customs. We also found the number of seasonal words with a highest frequency on the day is widely distributed and in the case of seasonal words with a specific day the distribution follows a power law. These findings would give support to writers about seasonal topics and to suggest seasonal items for shop staff.