Paper: arXiv 2402.04740

Authors: Sobin Joseph, Shashi Jain

Abstract

An extension of the Hawkes process, the Marked Hawkes process distinguishes itself by featuring variable jump size across each event, in contrast to the constant jump size observed in a Hawkes process without marks. While extensive literature has been dedicated to the non-parametric estimation of both the linear and non-linear Hawkes process, there remains a significant gap in the literature regarding the marked Hawkes process. In response to this, we propose a methodology for estimating the conditional intensity of the marked Hawkes process. We introduce two distinct models: \textit{“Shallow Neural Hawkes with marks”}- for Hawkes processes with excitatory kernels and \textit{“Neural Network for Non-Linear Hawkes with Marks”}- for non-linear Hawkes processes. Both these approaches take the past arrival times and their corresponding marks as the input to obtain the arrival intensity. This approach is entirely non-parametric, preserving the interpretability associated with the marked Hawkes process. To validate the efficacy of our method, we subject the method to synthetic datasets with known ground truth. Additionally, we apply our method to model cryptocurrency order book data, demonstrating its applicability to real-world scenarios.

Complexity vs Empirical Score

  • Math Complexity: 8.5/10
  • Empirical Rigor: 6.0/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents advanced mathematics including non-parametric estimation, neural network-based kernel approximation, and multi-dimensional point processes, while providing empirical validation on both synthetic and real cryptocurrency order book data.

Research Flowchart

  flowchart TD
  G["Research Goal: Non-parametric estimation for Marked Hawkes Processes"] --> I["Input Data<br>Synthetic Datasets (Ground Truth)<br>Cryptocurrency LOB Data (Past Times & Marks)"]
  I --> M["Key Methodology: Neural Network Architectures"]
  M --> M1["Model 1: Shallow Neural Hawkes with Marks<br>(For Excitatory Kernels)"]
  M --> M2["Model 2: Neural Network for Non-Linear Hawkes with Marks<br>(For Non-Linear Processes)"]
  M1 --> P["Computational Process<br>Non-parametric Estimation of Conditional Intensity<br>Utilizing Past Times & Marks"]
  M2 --> P
  P --> F["Key Findings & Outcomes<br>Validated efficacy on synthetic data<br>Demonstrated applicability on cryptocurrency LOB data<br>Bridged gap in Marked Hawkes literature"]

Evidence Card

What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card

Proposes

The paper proposes a non‑parametric methodology to estimate the conditional intensity of marked Hawkes processes, introducing two neural network models: Shallow Neural Hawkes with marks for excitatory kernels and Neural Network for Non‑Linear Hawkes with Marks for non‑linear kernels.

Source sentence
We propose a methodology for estimating the conditional intensity of the marked Hawkes process. We introduce two distinct models: Shallow Neural Hawkes with marks- for Hawkes processes with excitatory kernels and Neural Network for Non-Linear Hawkes with Marks- for non-linear Hawkes processes.

Assumes 2

  • Marks within each dimension are independent of marks in other dimensions and are i.i.d. statistical
    Source sentence
    our specific assumption is that marks within a dimension are independent of those in other dimensions.
  • The model accommodates continuous marks via a two‑layer feed‑forward network. model
    Source sentence
    The two-layered feed-forward network accommodates continuous marks, enabling a more comprehensive modelling of the marked Hawkes process.

Data

  • Synthetic datasets with known ground truth synthetic unknown
    Source sentence
    We subject the method to synthetic datasets with known ground truth.
  • Cryptocurrency market order book data (BTC‑USD and ETH‑USD) · Binance exchange order-book public
    16th July 2022, 02:00:00 UTC to 02:30:00 · Bitcoin and Ethereum market orders
    Source sentence
    The dataset comprises the arrival times and their associated volumes for market orders in Bitcoin‑US Dollar and Ethereum‑US Dollar pairs.

Recognised sources: Binance

Costs

Not applicable The paper reports no trading or execution result.

Finds 3

  • The estimated kernel closely matches the theoretical kernel for a 2‑dimensional marked Hawkes process, with low error shown in visual and quantitative plots. statistical
    Source sentence
    Figure 7: (a) The theoretical and the estimated kernel obtained using SNH with marks for a 2 −d marked Hawkes process with decoupled kernel function. (b) illustrates the error between the theoretical and estimated kernels
  • Quantile‑Quantile plots demonstrate high precision of inter‑arrival time predictions for the multivariate SNH and SNH with marks estimations. predictive-accuracy
    Source sentence
    Figure 8: QQ plot demonstrating the precision of prediction for multivariate SNH and SNH with marks estimation.
  • In the cryptocurrency application, estimation with marks uncovers the relationship between market order arrival intensity and trade volume, providing insight into market microstructure dynamics. other
    Source sentence
    estimation with marks reveals the relationship of the market order arrival intensity with the volume of the trade apart in addition to the history of trade arrivals.

Checked by others

No other paper in the archive has been found to support or contradict these findings (compared with 15 related papers). Coverage is partial, so this is neither replication nor refutation.

Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.