Hawkes Processes with Stochastic Excitations

Authors: Young Lee, Kar Wai Lim, Cheng Soon Ong

ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our sampling procedure scales linearly with the number of required events and does not require stationarity of the point process. A modular inference procedure consisting of a combination between Gibbs and Metropolis Hastings steps is put forward. We recover expectation maximization as a special case. Our general approach is illustrated for contagion following geometric Brownian motion and exponential Langevin dynamics. Figure 2. Versatility of Stochastic Hawkes: Observe that by allowing the level of contagion Y to be a stochastic process satisfying for instance a Geometric Brownian Motion (GBM) makes it possible to reproduce stylized facts of both ground truths of GBM and iid Gamma variates, see plots (a) & (b) and (d) & (f).
Researcher Affiliation Collaboration Young Lee YOUNG.LEE@NICTA.COM.AU Kar Wai Lim KARWAI.LIM@ANU.EDU.AU Cheng Soon Ong CHENGSOON.ONG@ANU.EDU.AU Data61/National ICT Australia & London School of Economics Data61/National ICT Australia & Australian National University
Pseudocode Yes Algorithm 1 Simulation of Stochastic Hawkes and Algorithm 2 MCMC Algorithm For Stochastic Hawkes
Open Source Code No The paper does not explicitly state that source code for the methodology described is provided or publicly available.
Open Datasets No The paper indicates using simulated data for illustrations and experiments (e.g., 'First we generate levels of contagion Y following a GBM'). It does not provide concrete access information (link, DOI, repository, or citation) for a publicly available or open dataset.
Dataset Splits No The paper describes simulation and inference but does not provide specific details on dataset splits (e.g., percentages, sample counts, or citations to predefined splits) for training, validation, or testing.
Hardware Specification No The paper does not specify any hardware details (e.g., CPU, GPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper discusses algorithms and statistical models but does not provide specific software dependencies (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup No The paper details the model specification and inference procedures, including the form of prior distributions for parameters (e.g., 'P(µ) N(µ0, σ2 0)'). However, it does not provide specific numerical values for these hyperparameters or other concrete training configurations (e.g., learning rates, batch sizes, or iterations) in the main text.