A Multitask Point Process Predictive Model

Wenzhao Lian, Ricardo Henao, Vinayak Rao, Joseph Lucas, Lawrence Carin
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:2030-2038, 2015.

Abstract

Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records.

Cite this Paper

BibTeX
@InProceedings{pmlr-v37-lian15, title = {A Multitask Point Process Predictive Model}, author = {Lian, Wenzhao and Henao, Ricardo and Rao, Vinayak and Lucas, Joseph and Carin, Lawrence}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {2030--2038}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/lian15.pdf}, url = {https://proceedings.mlr.press/v37/lian15.html}, abstract = {Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records.} }
Endnote
%0 Conference Paper %T A Multitask Point Process Predictive Model %A Wenzhao Lian %A Ricardo Henao %A Vinayak Rao %A Joseph Lucas %A Lawrence Carin %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-lian15 %I PMLR %P 2030--2038 %U https://proceedings.mlr.press/v37/lian15.html %V 37 %X Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records.
RIS
TY - CPAPER TI - A Multitask Point Process Predictive Model AU - Wenzhao Lian AU - Ricardo Henao AU - Vinayak Rao AU - Joseph Lucas AU - Lawrence Carin BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-lian15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 2030 EP - 2038 L1 - http://proceedings.mlr.press/v37/lian15.pdf UR - https://proceedings.mlr.press/v37/lian15.html AB - Point process data are commonly observed in fields like healthcare and social science. Designing predictive models for such event streams is an under-explored problem, due to often scarce training data. In this work we propose a multitask point process model, leveraging information from all tasks via a hierarchical Gaussian process (GP). Nonparametric learning functions implemented by a GP, which map from past events to future rates, allow analysis of flexible arrival patterns. To facilitate efficient inference, we propose a sparse construction for this hierarchical model, and derive a variational Bayes method for learning and inference. Experimental results are shown on both synthetic data and an application on real electronic health records. ER -
APA
Lian, W., Henao, R., Rao, V., Lucas, J. & Carin, L.. (2015). A Multitask Point Process Predictive Model. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:2030-2038 Available from https://proceedings.mlr.press/v37/lian15.html.

Related Material