Marked point processes and intensity ratios for limit order book modeling

This paper extends the analysis of Muni Toke and Yoshida (2020) to the case of marked point processes. We consider multiple marked point processes with intensities defined by three multiplicative components, namely a common baseline intensity, a state-dependent component specific to each process, and a state-dependent component specific to each mark within each process. We show that for specific mark distributions, this model is a combination of the ratio models defined in Muni Toke and Yoshida (2020). We prove convergence results for the quasi-maximum and quasi-Bayesian likelihood estimators of this model and provide numerical illustrations of the asymptotic variances. We use these ratio processes to model transactions occurring in a limit order book. Model flexibility allows us to investigate both state-dependency (emphasizing the role of imbalance and spread as significant signals) and clustering. Calibration, model selection and prediction results are reported for high-frequency trading data on multiple stocks traded on Euronext Paris. We show that the marked ratio model outperforms other intensity-based methods (such as “pure” Hawkes-based methods) in predicting the sign and aggressiveness of market orders on financial markets.

Extensions of Markov Modulated Poisson Processes and Their Applications to Deep Earthquakes

<p>The focus of this thesis is on the Markov modulated Poisson process (MMPP) and its extensions, aiming to propose appropriate statistical models for the occurrence patterns of main New Zealand deep earthquakes. Such an attempt might be beyond the scope of the MMPP and its extensions, however we hope its main patterns can be characterized by current models proposed in three parts of the thesis. The first part of the thesis is concerned with introductions and preliminaries of discrete time hidden Markov models (HMMs) and MMPP. The exibility in model formulation and openness in model framework of HMMs are reviewed in this part, suggesting also possible extensions of MMPP. The second part of the thesis is mainly about several extensions of MMPP. One extension of MMPP is by associating each occurrence of MMPP with a mark. Such an extension is potentially useful for spatial-temporal modelling or other point processes with marks. A special case of this type of extension is by allowing the multiple observations of MMPP synchronized together under the same Markov chain. This extension opens the possibility of modelling multiple point process observations with weak dependence. The third extension is motivated by the attempt to describe small scale temporal clustering existing in the deep earthquakes via treating the recognized aftershocks as marks which itself forms a finite point process. The rest of the second part focuses on some information theoretical aspects of MMPPs such as the entropy rate of the underlying Markov chain and observed point process respectively and their mutual information rate. A conjecture on the possible links between mutual information rate of MMPP and the Fisher information of the estimated parameters is suggested. The second part on extensions of MMPP is featured by the derivation of the likelihood and complete likelihood, parameter estimation via EM algorithm, state smoothing estimation and model evaluation through systematic applications of rescaling theory of multivariate point processes and marked point processes. The third part of the thesis includes the applications of these methods to the deep earthquakes in New Zealand. We first evaluate the data coverage, catalogue completeness and explore its descriptive characteristics and empirical properties such as epicentral distributions, depth distributions and magnitude distributions. Clustering behavior is studied via the second order moment analysis of point processes in the chapter 8. We also apply, the stress release models and the ETAS models which are usually used for shallow earthquakes, to the New Zealand deep earthquakes and provide tentative explanations of why they are not satisfactory for the deep earth-quakes. The chapter 9 is on the applications of MMPP and its extensions to the New Zealand deep earthquakes. Conclusions and future studies are presented in chapter 10.</p>

Extensions of Markov Modulated Poisson Processes and Their Applications to Deep Earthquakes

<p>The focus of this thesis is on the Markov modulated Poisson process (MMPP) and its extensions, aiming to propose appropriate statistical models for the occurrence patterns of main New Zealand deep earthquakes. Such an attempt might be beyond the scope of the MMPP and its extensions, however we hope its main patterns can be characterized by current models proposed in three parts of the thesis. The first part of the thesis is concerned with introductions and preliminaries of discrete time hidden Markov models (HMMs) and MMPP. The exibility in model formulation and openness in model framework of HMMs are reviewed in this part, suggesting also possible extensions of MMPP. The second part of the thesis is mainly about several extensions of MMPP. One extension of MMPP is by associating each occurrence of MMPP with a mark. Such an extension is potentially useful for spatial-temporal modelling or other point processes with marks. A special case of this type of extension is by allowing the multiple observations of MMPP synchronized together under the same Markov chain. This extension opens the possibility of modelling multiple point process observations with weak dependence. The third extension is motivated by the attempt to describe small scale temporal clustering existing in the deep earthquakes via treating the recognized aftershocks as marks which itself forms a finite point process. The rest of the second part focuses on some information theoretical aspects of MMPPs such as the entropy rate of the underlying Markov chain and observed point process respectively and their mutual information rate. A conjecture on the possible links between mutual information rate of MMPP and the Fisher information of the estimated parameters is suggested. The second part on extensions of MMPP is featured by the derivation of the likelihood and complete likelihood, parameter estimation via EM algorithm, state smoothing estimation and model evaluation through systematic applications of rescaling theory of multivariate point processes and marked point processes. The third part of the thesis includes the applications of these methods to the deep earthquakes in New Zealand. We first evaluate the data coverage, catalogue completeness and explore its descriptive characteristics and empirical properties such as epicentral distributions, depth distributions and magnitude distributions. Clustering behavior is studied via the second order moment analysis of point processes in the chapter 8. We also apply, the stress release models and the ETAS models which are usually used for shallow earthquakes, to the New Zealand deep earthquakes and provide tentative explanations of why they are not satisfactory for the deep earth-quakes. The chapter 9 is on the applications of MMPP and its extensions to the New Zealand deep earthquakes. Conclusions and future studies are presented in chapter 10.</p>

A comprehensive model for cyber risk based on marked point processes and its application to insurance

After scrutinizing technical, legal, financial, and actuarial aspects of cyber risk, a new approach for modelling cyber risk using marked point processes is proposed. Key covariates, required to model frequency and severity of cyber claims, are identified. The presented framework explicitly takes into account incidents from malicious untargeted and targeted attacks as well as accidents and failures. The resulting model is able to include the dynamic nature of cyber risk, while capturing accumulation risk in a realistic way. The model is studied with respect to its statistical properties and applied to the pricing of cyber insurance and risk measurement. The results are illustrated in a simulation study.

Time-dynamic evaluations under non-monotone information generated by marked point processes

The information dynamics in finance and insurance applications is usually modelled by a filtration. This paper looks at situations where information restrictions apply so that the information dynamics may become non-monotone. A fundamental tool for calculating and managing risks in finance and insurance are martingale representations. We present a general theory that extends classical martingale representations to non-monotone information generated by marked point processes. The central idea is to focus only on those properties that martingales and compensators show on infinitesimally short intervals. While classical martingale representations describe innovations only, our representations have an additional symmetric counterpart that quantifies the effect of information loss. We exemplify the results with examples from life insurance and credit risk.