[4A2] 機械学習-深層学習(6)

[4A2-05] Newcomers churn prediction by the event sequence on social networking service

〇Koya Sato¹, Mizuki Oka¹, Kazuhiko Kato¹ (1. Tsukuba University)

キーワード：Churn prediction, Deep Learning, SNS

To be sustainable the social networking service, newcomer churn prediction is an important task. Previous churn prediction models use handcrafted and service-specific features. Therefore, it is difficult to apply developed the model to other similar social networking services. To solve this, we propose new deep learning method that doesn't depend on specific service.