Home // International Journal On Advances in Internet Technology, volume 10, numbers 1 and 2, 2017 // View article

Detection of Japanese and English Tweets Where Birthdays are Revealed to Other People

Authors:
Yasuhiko Watanabe
Naohiro Miyagi
Kenji Yasuda
Norimasa Mukai
Ryo Nishimura
Yoshihiro Okada

Keywords: birthday, personal information, Twitter, SNS, privacy risk

Abstract:
These days, many people use a social networking service (SNS). When we use SNSs, we carefully protect the privacy of personal information: name, age, gender, address, telephone number, birthday, etc. However, we often reveal birthdays on SNS, not only ours but also of others. Birthday information can threaten our privacy and security when combined with other personal information. In this study, we investigated Japanese and English tweets where birthdays were revealed to other people, including unwanted audiences. We collected 1,000 Japanese tweets and 1,000 English tweets including word ``birthday'' and found about 30% of the collected Japanese tweets and 70% of the English tweets were tweets revealing someone's birthdays to other people. Furthermore, about 70% of Japanese tweets and 90% of English tweets revealing someone's birthdays to other people were ones where receivers' birthdays were revealed. We obtained 88% accuracy when we applied support vector machine (SVM) machine learning techniques to classify Japanese and English tweets including word ``birthday'' into ones revealing birthdays of senders, receivers, and others. However, the recall rate of Japanese and English tweets revealing senders' birthdays were only 17% and 30%, respectively.

Pages: 87 to 96

Copyright: Copyright (c) to authors, 2017. Used with permission.

Publication date: June 30, 2017

Published in: journal

ISSN: 1942-2652