Efficient Non-parametric Bayesian Hawkes Processes

@inproceedings{Zhang2019EfficientNB,
  title={Efficient Non-parametric Bayesian Hawkes Processes},
  author={Rui Zhang and Christian J. Walder and Marian-Andrei Rizoiu and Lexing Xie},
  booktitle={IJCAI},
  year={2019}
}
In this paper, we develop an efficient non-parametric Bayesian estimation of the kernel function of Hawkes processes. The non-parametric Bayesian approach is important because it provides flexible Hawkes kernels and quantifies their uncertainty. Our method is based on the cluster representation of Hawkes processes. Utilizing the stationarity of the Hawkes process, we efficiently sample random branching structures and thus, we split the Hawkes process into clusters of Poisson processes. We… 

Figures and Tables from this paper

Sparse Gaussian Process Modulated Hawkes Process
TLDR
This paper proposes a new non-parametric Bayesian Hawkes process whose triggering kernel is modeled as a squared sparse Gaussian process and presents the variational inference scheme for the model optimization, which has the advantage of linear time complexity by leveraging the stationarity of the triggering kernel.
Efficient Inference for Nonparametric Hawkes Processes Using Auxiliary Latent Variables
TLDR
The baseline intensity and triggering kernel are both modeled as the sigmoid transformation of random trajectories drawn from a GP as well as an efficient mean-field variational inference algorithm to approximate the posterior and a sparse GP approximation is introduced to reduce complexity.
Variational Bayesian Inference for Nonlinear Hawkes Process with Gaussian Process Self-Effects
TLDR
This work proposes an extended model where the self-effects are of both excitatory and inhibitory types and follow a Gaussian Process, and describes a mean-field variational inference approach to learn the model parameters.
Nonlinear Hawkes Processes in Time-Varying System
TLDR
This work aims to overcome all three assumptions simultaneously by proposing the flexible state-switching Hawkes processes: a flexible, nonlinear and nonhomogeneous variant where a state process is incorporated to interact with the point processes.
Deep Neyman-Scott Processes
TLDR
A deep Neyman-Scott process is considered, for which the building components of a network are all Poisson processes, and an e-cient posterior sampling via Markov chain Monte Carlo is developed, which opens up room for the inference in sophisti-cated hierarchical point processes.
Interval-censored Hawkes processes
TLDR
The Mean Behavior Poisson (MBP) process is defined, a novel Poisson process with a direct parameter correspondence to the popular self-exciting Hawkes process and it is shown that on real-world datasets that the MBP process outperforms HIP for the task of popularity prediction.
GP-ETAS: semiparametric Bayesian inference for the spatio-temporal epidemic type aftershock sequence model
TLDR
This work introduces a highly flexible, non-parametric representation for the spatially varying ETAS background intensity through a Gaussian process (GP) prior and demonstrates the predictive power for observed earthquake catalogues including uncertainty quantification for the estimated parameters.
THP: Topological Hawkes Processes for Learning Granger Causality on Event Sequences
TLDR
A Topological Hawkes processes (THP) is proposed to draw a connection between the graph convolution in topology domain and the temporal Convolution in time domains and the Granger causality learning method on THP in a likelihood framework is proposed.
THPs: Topological Hawkes Processes for Learning Causal Structure on Event Sequences.
TLDR
A topological Hawkes process (THP) is proposed to draw a connection between the graph convolution in the topology domain and the temporal Convolution in time domains to learn causal structure learning on THP in a likelihood framework.
Flexible Temporal Point Processes Modeling with Nonlinear Hawkes Processes with Gaussian Processes Excitations and Inhibitions
TLDR
An extended Hawkes process model where the self–effects are of both excitatory and inhibitory type and follow a Gaussian Process is proposed, and efficient approximate Bayesian inference is achieved via data augmentation, and a mean–field variational inference approach is described to learn the model parameters.
...
...

References

SHOWING 1-10 OF 71 REFERENCES
Fast and scalable non-parametric Bayesian inference for Poisson point processes
TLDR
It is shown that the second approach depends in a less sensitive way on the choice of the number of bins $N$ and outperforms the first approach in practice, and also scales extremely well with the amount of data.
An estimation procedure for the Hawkes process
In this paper, we present a nonparametric estimation procedure for the multivariate Hawkes point process. The timeline is cut into bins and—for each component process—the number of points in each bin
Nonparametric Bayesian estimation for multivariate Hawkes processes
TLDR
This paper studies nonparametric estimation of parameters of multivariate Hawkes processes by considering the Bayesian setting and derive posterior concentration rates from L1-metrics for stochastic intensities of the Hawkes process and using priors based on piecewise constant functions.
Scalable Nonparametric Bayesian Inference on Point Processes with Gaussian Processes
TLDR
An efficient, scalable non-parametric Gaussian process model for inference on Poisson point processes is proposed and it is shown that the model obtained is faster, more accurate and generates less correlated samples than competing approaches on both synthetic and real-life data.
Bayesian Inference for Hawkes Processes
TLDR
This paper explores and compares two approaches to Bayesian inference based on an underlying clustering and branching structure in the Hawkes process and compares them numerically.
RESEARCH ARTICLE A Nonparametric EM algorithm for Multiscale Hawkes Processes
TLDR
Maximum penalized likelihood estimation is proposed as a method for simultaneously estimating the background rate and the triggering density of Hawkes process intensities that vary over multiple time scales and used to examine self-excitation in Iraq IED event patterns.
First- and Second-Order Statistics Characterization of Hawkes Processes and Non-Parametric Estimation
TLDR
A systematic study of a fast and efficient method, which main principles were initially sketched by Bacry and Muzy, to perform a non-parametric estimation of the Hawkes kernel matrix in the general framework of marked Hawkes processes.
Tractable nonparametric Bayesian inference in Poisson processes with Gaussian process intensities
TLDR
This paper presents the first approach to Gaussian Cox processes in which it is possible to perform inference without introducing approximations or finitedimensional proxy distributions, and uses a generative model for Poisson data to enable tractable inference via Markov chain Monte Carlo.
Fast Bayesian Intensity Estimation for the Permanental Process
TLDR
This paper presents a fast Bayesian inference scheme for the permanental process, a Cox process under which the square root of the intensity is a Gaussian process, and exploits connections with reproducing kernel Hilbert spaces to derive efficient approximate Bayes inference algorithms.
Learning Granger Causality for Hawkes Processes
TLDR
An effective learning algorithm combining a maximum likelihood estimator (MLE) with a sparse-group-lasso (SGL) regularizer is proposed that can learn the Granger causality graph and the triggering patterns of the Hawkes processes simultaneously.
...
...