Increasing users' engagement by providing personalized notifications in private social networks for healthcare

在医疗保健的私人社交网络中，患者之间丰富的社交互动加强了 网络提供情感支持和社区感的能力。这种社区和同龄人的感觉 支持对于库拉蒂奥提供的私人社交网络的成功至关重要。为了增加这种感觉 社区我们需要及时主动地通知用户网络中的新事件和互动 使得当网络中的任何患者发起社交交互时，他/她在网络中的同龄人可以 在合理的时间框架内对此做出回应。此外，随着网络规模的扩大，医疗从业者 在这些网络中扮演支持角色的人将从智能通知系统中受益匪浅 可以引起他们注意网络中在医学方面具有重要意义的帖子或消息，以及 需要他们的及时关注。 Curatio提供的社交网络产品是在移动平台上呈现的，这为我们提供了 有机会利用本平台提供的“推送通知”机制鼓励用户参与 更多的时候是与网络上的其他用户进行社交互动。通知越来越多地被用于各种 争相吸引用户注意力和时间的移动应用程序。然而，很少有研究是 根据这些通知的成功率提供。我们认为有必要管理相关性、质量、 发送给用户的通知的频率和时间，以增加 每个用户的通知。该项目旨在应用机器学习技术来学习用户模型， 通过利用过去的活动数据以及每个用户的 对S/他从系统收到的通知的反应。

In a private social network for healthcare, abundance of social interactions between patients strengthens the network's ability to provide emotional support and sense of community. This sense of community and peer support is essential for the success of the private social networks provided by Curatio. To increase this sense of community we need to proactively notify users of new events and interactions in the network in a timely manner so that when any patient in the network initiates a social interaction, his/her peers in the network can respond to it in a reasonable time frame. Additionally, as the networks grow in size, the healthcare practitioners who play the support role in these networks would greatly benefit from an intelligent notification system that could bring to their attention the posts or messages in the network that are significant in medical terms and needs their prompt attention. The social network products provided by Curatio are presented in a mobile platform, which gives us the opportunity to use the "push notification" mechanism provided in this platform for encouraging users to engage more often in social interactions with other users on the network. Notifications are increasingly used in various mobile applications which compete for attracting the user's attention and time. However, little research is available on success rate of these notifications. We believe it is necessary to manage the relevance, quality, frequency, and timing of the notifications sent to the user to increase the effectiveness and usefulness of the notifications for each user. This project aims to apply machine learning techniques to learn user models that personalize the above mentioned factors to each user, by utilizing the past activity data as well as each user's reactions to the notifications s/he receives from the system.