Loading...

User-Centric Recommendation for Mobile Notification Servicees

Jami Moghaddam, Iman | 2019

1063 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 52500 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Soleymani Baghshah, Mahdieh
  7. Abstract:
  8. With the popularity of smart devices, a lot of applications have developed and deployed.Developers try to establish continuous interaction with their users by different tools including push notifications. Push notification is a message that is sent from developers to the users and as soon as the user’s device receives that, it appears on the device screen. Sending proper content to users in order to resume their engagement is one of the most important usages of notifications. Users are not interested in receiving irrelevant notifications, and receiving irrelevant notifications make them remove the application, so it’s important to predict users’ interest in different notifications and push the notification only to the users that click on it with high probability.According to recommender systems definition, predicting the users’ interest in notifications is a recommendation problem and we call it ”notification recommendation”. In this study, we face some particular recommendation challenges, such as cold start and changing users’ interest over time. According to Deep Neural Networks Successes in the field of recommendation, we present deep neural networks based methods to solve notification recommendation problem. Proposed methods in this study are presented by three major approaches. The first approach that is based on multilayer perceptron networks, tries to model static features of users interest in different notifications. Methods that are proposed by the second approach are based on recurrent neural networks and models change of users’interest over time. Eventually, the third approach uses static and dynamic features together by using recurrent and mutilayer perceptron networks simultaneously. Proposed methods are evaluated with Pushe service push notifications data set by precision and recall metrics. By applying the proposed recurrent neural network based method, the best results were achieved.Using this method for sending notifications to users resulted in 0.125 precision at 0.6 recall,while broadcasting the notification to all users, resulted in 0.045 precision
  9. Keywords:
  10. Deep Neural Networks ; Recommender System ; Push Notification ; Mobile Push Notification Service

 Digital Object List

 Bookmark

...see more