Multivariate Count Data Models for Time Series Forecasting
Count data appears in many research fields and exhibits certain features that make modeling difficult. Most popular approaches to modeling count data can be classified into observation and parameter-driven models. In this paper, we review two models from these classes: the log-linear multivariate co...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-06-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/23/6/718 |
| _version_ | 1797531328754745344 |
|---|---|
| author | Yuliya Shapovalova Nalan Baştürk Michael Eichler |
| author_facet | Yuliya Shapovalova Nalan Baştürk Michael Eichler |
| author_sort | Yuliya Shapovalova |
| collection | DOAJ |
| description | Count data appears in many research fields and exhibits certain features that make modeling difficult. Most popular approaches to modeling count data can be classified into observation and parameter-driven models. In this paper, we review two models from these classes: the log-linear multivariate conditional intensity model (also referred to as an integer-valued generalized autoregressive conditional heteroskedastic model) and the non-linear state-space model for count data. We compare these models in terms of forecasting performance on simulated data and two real datasets. In simulations, we consider the case of model misspecification. We find that both models have advantages in different situations, and we discuss the pros and cons of inference for both models in detail. |
| first_indexed | 2024-03-10T10:42:13Z |
| format | Article |
| id | doaj.art-a5c5f6da0b9b42d5868b9d2a9f98688c |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-03-10T10:42:13Z |
| publishDate | 2021-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-a5c5f6da0b9b42d5868b9d2a9f98688c2023-11-21T22:54:06ZengMDPI AGEntropy1099-43002021-06-0123671810.3390/e23060718Multivariate Count Data Models for Time Series ForecastingYuliya Shapovalova0Nalan Baştürk1Michael Eichler2Institute for Computing and Information Sciences, Radboud University Nijmegen, Toernooiveld 212, 6525 EC Nijmegen, The NetherlandsSchool of Business and Economics, Maastricht University, Tongersestraat 53, 6211 LM Maastricht, The NetherlandsSchool of Business and Economics, Maastricht University, Tongersestraat 53, 6211 LM Maastricht, The NetherlandsCount data appears in many research fields and exhibits certain features that make modeling difficult. Most popular approaches to modeling count data can be classified into observation and parameter-driven models. In this paper, we review two models from these classes: the log-linear multivariate conditional intensity model (also referred to as an integer-valued generalized autoregressive conditional heteroskedastic model) and the non-linear state-space model for count data. We compare these models in terms of forecasting performance on simulated data and two real datasets. In simulations, we consider the case of model misspecification. We find that both models have advantages in different situations, and we discuss the pros and cons of inference for both models in detail.https://www.mdpi.com/1099-4300/23/6/718multivariate count dataINGACRCHstate-space modelbank failurestransactions |
| spellingShingle | Yuliya Shapovalova Nalan Baştürk Michael Eichler Multivariate Count Data Models for Time Series Forecasting Entropy multivariate count data INGACRCH state-space model bank failures transactions |
| title | Multivariate Count Data Models for Time Series Forecasting |
| title_full | Multivariate Count Data Models for Time Series Forecasting |
| title_fullStr | Multivariate Count Data Models for Time Series Forecasting |
| title_full_unstemmed | Multivariate Count Data Models for Time Series Forecasting |
| title_short | Multivariate Count Data Models for Time Series Forecasting |
| title_sort | multivariate count data models for time series forecasting |
| topic | multivariate count data INGACRCH state-space model bank failures transactions |
| url | https://www.mdpi.com/1099-4300/23/6/718 |
| work_keys_str_mv | AT yuliyashapovalova multivariatecountdatamodelsfortimeseriesforecasting AT nalanbasturk multivariatecountdatamodelsfortimeseriesforecasting AT michaeleichler multivariatecountdatamodelsfortimeseriesforecasting |