Subtopic Deep Dive
Hawkes Processes
Research Guide
What is Hawkes Processes?
Hawkes processes are self-exciting point processes where the intensity function increases with each event according to an excitation kernel.
Introduced by Alan G. Hawkes in 1971, these models capture contagion effects in event sequences like earthquakes and financial trades. Key developments include maximum likelihood estimation by Ozaki (1979, 313 citations) and stability analysis for nonlinear extensions by Brémaud and Massoulié (1996, 417 citations). Applications span finance (Bacry et al., 2015, 390 citations) and social networks (Zhou et al., 2013, 295 citations).
Why It Matters
Hawkes processes model clustering in financial order books, enabling resiliency measurement (Large, 2006, 277 citations) and value-at-risk estimation (Chavez-Demoulin et al., 2005, 191 citations). In multivariate settings, they quantify cross-excitations in market data via maximum likelihood (Embrechts et al., 2011, 270 citations). Social infectivity estimation uses low-rank network adaptations (Zhou et al., 2013, 295 citations), informing contagion forecasting across seismology, neuroscience, and epidemics.
Key Research Challenges
Nonlinear Stability Analysis
Nonlinear Hawkes processes lack guaranteed convergence to equilibrium unlike linear cases. Brémaud and Massoulié (1996, 417 citations) provide conditions for stability but extensions to high dimensions remain open. Computational verification of branching ratios exceeds critical thresholds is needed.
Kernel Parameter Estimation
Maximum likelihood estimation suffers from identifiability issues in multivariate settings (Ozaki, 1979, 313 citations). Embrechts et al. (2011, 270 citations) derive estimators but goodness-of-fit tests require innovation. High-frequency financial data amplifies overfitting risks.
Multivariate Cross-Excitation
Capturing asymmetric influences between event types demands complex kernel structures (Bacry et al., 2015, 390 citations). Limit theorems for statistics like spectral analysis are limited (Bacry et al., 2013, 242 citations). Scalability to sparse networks challenges inference (Zhou et al., 2013, 295 citations).
Essential Papers
Stability of nonlinear Hawkes processes
Pierre Brémaud, Laurent Massoulié · 1996 · The Annals of Probability · 417 citations
We address the problem of the convergence to equilibrium of a\ngeneral class of point processes, containing, in particular, the nonlinear\nmutually exciting point processes, an extension of the lin...
Hawkes Processes in Finance
Emmanuel Bacry, Iacopo Mastromatteo, Jean–François Muzy · 2015 · Market Microstructure and Liquidity · 390 citations
In this paper we propose an overview of the recent academic literature devoted to the applications of Hawkes processes in finance. Hawkes processes constitute a particular class of multivariate poi...
Maximum likelihood estimation of Hawkes' self-exciting point processes
T. Ozaki · 1979 · Annals of the Institute of Statistical Mathematics · 313 citations
Learning Social Infectivity in Sparse Low-rank Networks Using Multi-dimensional Hawkes Processes
Ke Zhou, Hongyuan Zha, Le Song · 2013 · 295 citations
How will the behaviors of individuals in a social network be influenced by their neigh-bors, the authorities and the communities in a quantitative way? Such critical and valu-able knowledge is unfo...
Measuring the resiliency of an electronic limit order book
Jeremy Large · 2006 · Journal of Financial Markets · 277 citations
Multivariate Hawkes processes: an application to financial data
Paul Embrechts, Thomas Liniger, Lu Lin · 2011 · Journal of Applied Probability · 270 citations
A Hawkes process is also known under the name of a self-exciting point process and has numerous applications throughout science and engineering. We derive the statistical estimation (maximum likeli...
Some limit theorems for Hawkes processes and application to financial statistics
Emmanuel Bacry, Sylvain Delattre, Marc Hoffmann et al. · 2013 · Stochastic Processes and their Applications · 242 citations
Reading Guide
Foundational Papers
Start with Hawkes (1971, 192 citations) for spectra and basic excitation; Ozaki (1979, 313 citations) for MLE; Brémaud and Massoulié (1996, 417 citations) for stability—these establish core theory and estimation.
Recent Advances
Bacry et al. (2015, 390 citations) for finance overview; Embrechts et al. (2011, 270 citations) for multivariate fitting; Zhou et al. (2013, 295 citations) for network adaptations.
Core Methods
Exponential/power-law kernels; maximum likelihood with contrast functions; spectral methods for branching ratio (Hawkes, 1971); cluster representation for simulation.
How PapersFlow Helps You Research Hawkes Processes
Discover & Search
Research Agent uses citationGraph on Brémaud and Massoulié (1996) to map stability lineages, then findSimilarPapers for nonlinear extensions. exaSearch queries 'Hawkes process earthquake clustering' retrieves 50+ applied papers. searchPapers with 'multivariate Hawkes finance' surfaces Embrechts et al. (2011).
Analyze & Verify
Analysis Agent runs readPaperContent on Bacry et al. (2015) to extract kernel forms, then verifyResponse with CoVe against Ozaki (1979) estimation claims. runPythonAnalysis simulates branching ratios with NumPy for stability checks, graded by GRADE for statistical rigor. Python sandbox fits multivariate models from Embrechts et al. (2011) data excerpts.
Synthesize & Write
Synthesis Agent detects gaps in nonlinear stability post-Brémaud (1996) via contradiction flagging across 20 papers. Writing Agent applies latexEditText to kernel equations, latexSyncCitations for Hawkes (1971), and latexCompile for publication-ready sections. exportMermaid diagrams excitation networks from Zhou et al. (2013).
Use Cases
"Simulate Hawkes process branching ratio exceeding 1 for instability"
Research Agent → searchPapers 'Hawkes stability Brémaud' → Analysis Agent → runPythonAnalysis (NumPy simulation of nonlinear intensity with exponential kernel) → matplotlib plot of divergence, outputting fitted parameters and stability verdict.
"Draft LaTeX section on multivariate Hawkes estimation with citations"
Synthesis Agent → gap detection in estimation literature → Writing Agent → latexEditText (insert Ozaki 1979 equations) → latexSyncCitations (Embrechts 2011) → latexCompile → PDF with compiled multivariate likelihood and financial example.
"Find GitHub code for Hawkes MLE in finance"
Research Agent → searchPapers 'Hawkes finance Bacry' → Code Discovery → paperExtractUrls → paperFindGithubRepo → githubRepoInspect → verified repo with Jupyter notebooks for order book simulation from Large (2006).
Automated Workflows
Deep Research workflow scans 50+ Hawkes papers via searchPapers → citationGraph → structured report on estimation evolution from Ozaki (1979) to Bacry (2015). DeepScan applies 7-step CoVe chain: readPaperContent (Embrechts 2011) → runPythonAnalysis replication → GRADE scoring. Theorizer generates hypotheses on spectral limits from Hawkes (1971) and Bacry et al. (2013).
Frequently Asked Questions
What defines a Hawkes process?
A Hawkes process has conditional intensity λ(t) = μ + ∫ α φ(t-s) dN(s), where μ is background rate, α excitation strength, φ kernel, and N counting process (Hawkes, 1971, 192 citations).
What are main estimation methods?
Maximum likelihood estimation solves ∫ log λ(t_i) dN(t_i) - ∫ λ(t) dt, with EM approximations for intractable kernels (Ozaki, 1979, 313 citations; Embrechts et al., 2011, 270 citations).
What are key papers?
Foundational: Hawkes (1971, spectra, 192 citations), Ozaki (1979, MLE, 313 citations), Brémaud and Massoulié (1996, stability, 417 citations). Recent: Bacry et al. (2015, finance review, 390 citations), Hawkes (2017, 218 citations).
What are open problems?
Scalable inference for high-dimensional nonlinear processes; robust estimation under model misspecification; theoretical limits for non-stationary excitations beyond Brémaud (1996).
Research Point processes and geometric inequalities with AI
PapersFlow provides specialized AI tools for Mathematics researchers. Here are the most relevant for this topic:
AI Literature Review
Automate paper discovery and synthesis across 474M+ papers
Paper Summarizer
Get structured summaries of any paper in seconds
AI Academic Writing
Write research papers with AI assistance and LaTeX support
See how researchers in Physics & Mathematics use PapersFlow
Field-specific workflows, example queries, and use cases.
Start Researching Hawkes Processes with AI
Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.
See how PapersFlow works for Mathematics researchers