Deep Representation Learning for Prediction of Temporal Event Sets in the Continuous Time Domain

Parag Dutta, Kawin Mayilvaghanan, Pratyaksha Sinha, Ambedkar Dukkipati
Proceedings of the 15th Asian Conference on Machine Learning, PMLR 222:343-358, 2024.

Abstract

Temporal Point Processes (TPP) play an important role in predicting or forecasting events. Although these problems have been studied extensively, predicting multiple simultaneously occurring events can be challenging. For instance, more often than not, a patient gets admitted to a hospital with multiple conditions at a time. Similarly people buy more than one stock and multiple news breaks out at the same time. Moreover, these events do not occur at discrete time intervals, and forecasting event sets in the continuous time domain remains an open problem. Naïve approaches for extending the existing TPP models for solving this problem lead to dealing with an exponentially large number of events or ignoring set dependencies among events. In this work, we propose a scalable and efficient approach based on TPPs to solve this problem. Our proposed approach incorporates contextual event embeddings, temporal information, and domain features to model the temporal event sets. We demonstrate the effectiveness of our approach through extensive experiments on multiple datasets, showing that our model outperforms existing methods in terms of prediction metrics and computational efficiency. To the best of our knowledge, this is the first work that solves the problem of predicting event set intensities in the continuous time domain by using TPPs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v222-dutta24a, title = {Deep Representation Learning for Prediction of Temporal Event Sets in the Continuous Time Domain}, author = {Dutta, Parag and Mayilvaghanan, Kawin and Sinha, Pratyaksha and Dukkipati, Ambedkar}, booktitle = {Proceedings of the 15th Asian Conference on Machine Learning}, pages = {343--358}, year = {2024}, editor = {Yanıkoğlu, Berrin and Buntine, Wray}, volume = {222}, series = {Proceedings of Machine Learning Research}, month = {11--14 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v222/dutta24a/dutta24a.pdf}, url = {https://proceedings.mlr.press/v222/dutta24a.html}, abstract = {Temporal Point Processes (TPP) play an important role in predicting or forecasting events. Although these problems have been studied extensively, predicting multiple simultaneously occurring events can be challenging. For instance, more often than not, a patient gets admitted to a hospital with multiple conditions at a time. Similarly people buy more than one stock and multiple news breaks out at the same time. Moreover, these events do not occur at discrete time intervals, and forecasting event sets in the continuous time domain remains an open problem. Naïve approaches for extending the existing TPP models for solving this problem lead to dealing with an exponentially large number of events or ignoring set dependencies among events. In this work, we propose a scalable and efficient approach based on TPPs to solve this problem. Our proposed approach incorporates contextual event embeddings, temporal information, and domain features to model the temporal event sets. We demonstrate the effectiveness of our approach through extensive experiments on multiple datasets, showing that our model outperforms existing methods in terms of prediction metrics and computational efficiency. To the best of our knowledge, this is the first work that solves the problem of predicting event set intensities in the continuous time domain by using TPPs.} }
Endnote
%0 Conference Paper %T Deep Representation Learning for Prediction of Temporal Event Sets in the Continuous Time Domain %A Parag Dutta %A Kawin Mayilvaghanan %A Pratyaksha Sinha %A Ambedkar Dukkipati %B Proceedings of the 15th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Berrin Yanıkoğlu %E Wray Buntine %F pmlr-v222-dutta24a %I PMLR %P 343--358 %U https://proceedings.mlr.press/v222/dutta24a.html %V 222 %X Temporal Point Processes (TPP) play an important role in predicting or forecasting events. Although these problems have been studied extensively, predicting multiple simultaneously occurring events can be challenging. For instance, more often than not, a patient gets admitted to a hospital with multiple conditions at a time. Similarly people buy more than one stock and multiple news breaks out at the same time. Moreover, these events do not occur at discrete time intervals, and forecasting event sets in the continuous time domain remains an open problem. Naïve approaches for extending the existing TPP models for solving this problem lead to dealing with an exponentially large number of events or ignoring set dependencies among events. In this work, we propose a scalable and efficient approach based on TPPs to solve this problem. Our proposed approach incorporates contextual event embeddings, temporal information, and domain features to model the temporal event sets. We demonstrate the effectiveness of our approach through extensive experiments on multiple datasets, showing that our model outperforms existing methods in terms of prediction metrics and computational efficiency. To the best of our knowledge, this is the first work that solves the problem of predicting event set intensities in the continuous time domain by using TPPs.
APA
Dutta, P., Mayilvaghanan, K., Sinha, P. & Dukkipati, A.. (2024). Deep Representation Learning for Prediction of Temporal Event Sets in the Continuous Time Domain. Proceedings of the 15th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 222:343-358 Available from https://proceedings.mlr.press/v222/dutta24a.html.

Related Material