Loading...

Continuous Time Modeling of Marked Events

Hosseini, Abbas | 2019

3057 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 51994 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Rabiee, Hamid Reza
  7. Abstract:
  8. A great deal of information are continuously generated by users in different contexts such as social networks and online service providers in terms of temporal marked events. These events indicate that what happened to who by when and where.Modeling such events and predicting future ones has interesting applications in different domains such as item recommendation in online service providers and trending topic prediction in online social networks. However, complex longitudinal dependencies among such events makes the prediction task challenging. Moreover, nonstationarity of generative model of events and large size of events, makes the modeling and learning the models challenging.In this thesis, we propose a unified statistical framework for joint modeling of mark and time and events which is able to learn the latent underlying patterns of events and their corresponding dynamics to predict the time and mark of future events. This framework, provide us a structured mechanism for modeling, inference and predicting events. Based on this structure, we propose three frameworks, i.e. Factorized Point Processes, Hierarchical Nonparametric Point Processes and Neural Point Processes. Each of these frameworks models the marked events using different tools and approaches. Moreover, in order to evaluate the performance of these frameworks, we proposed different methods based on these frameworks in different applications such as time-sensitive recommendation, content diffusion over social networks and churn prediction and compared the results with state of the art in each application domain
  9. Keywords:
  10. Marked Events ; Temporal Point Process ; Poisson Factorization ; Dependent Bayesian Nonparametric Models ; Time-Sensitive Recommender Systems ; Sequential Monte Carlo Method ; Content Diffusion ; Bayesian Variational Inference ; Variational Recurrent Neural Networks

 Digital Object List

 Bookmark

...see more