Abstract:In this paper, a time-varying first-order mixture integer-valued threshold autoregressive process driven by explanatory variables is introduced. The basic probabilistic and statistical properties of this model are studied in depth. We proceed to derive estimators using the conditional least squares (CLS) and conditional maximum likelihood (CML) methods, while also establishing the asymptotic properties of the CLS estimator. Furthermore, we employed the CLS and CML score functions to infer the threshold parameter. Additionally, three test statistics to detect the existence of the piecewise structure and explanatory variables were utilized. To support our findings, we conducted simulation studies and applied our model to two applications concerning the daily stock trading volumes of VOW.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to construct a Time - Varying Mixture Integer - Valued Threshold Autoregressive Process (TVMTTINAR(1)) that can effectively handle nonlinear phenomena and the influence of external variables in integer - valued time - series data. Specifically, the paper mainly focuses on the following aspects:
1. **Nonlinear phenomena**:
- The paper points out that traditional integer autoregressive models (INAR models) often perform poorly when dealing with nonlinear phenomena. For example, in epidemiology, the incidence rate of an epidemic may fluctuate over time, and using traditional INAR models may not be able to accurately capture such fluctuations.
- To deal with these nonlinear phenomena, the paper introduces the Time - Varying Mixture Integer - Valued Threshold Autoregressive Process (TVMTTINAR(1)), which can better describe the nonlinear structure in time series.
2. **The influence of external variables**:
- The paper emphasizes that integer - valued time - series data are usually affected by external variables. For example, the daily trading volume of stocks is not only affected by its own historical data, but may also be affected by market factors, industry events, interest rate policies, political and geopolitical factors, etc.
- Therefore, the paper proposes a TVMTTINAR(1) model driven by explanatory variables to more accurately capture the influence of these external variables on time - series data.
3. **Statistical inference of the model**:
- The paper studies the basic probability and statistical properties of the TVMTTINAR(1) model and proposes parameter estimation methods based on the Conditional Least Squares (CLS) and Conditional Maximum Likelihood (CML).
- The paper also establishes the asymptotic properties of the CLS estimator and uses the CLS and CML score functions to infer the threshold parameters.
4. **Model verification**:
- To verify the effectiveness of the model, the paper conducts simulation studies and applies the model to two actual data sets, namely the daily trading volume data of a certain stock.
- Through these studies, the paper shows the competitiveness of the proposed TVMTTINAR(1) model in dealing with integer - valued time - series data.
In summary, the main objective of this paper is to develop an integer - valued time - series model that can handle nonlinear phenomena and the influence of external variables and provide effective statistical inference methods to improve the modeling and analysis ability of integer - valued time - series data.