Distributed lags using elastic‑net regularization for market response models: focus on predictive and explanatory capacity
Over the course of several decades, extensive research has been dedicated to Market Response models, often lacking validation in purely predictive tasks and frequently overlooking the adherence of methods to underlying assumptions and conditions, such as the capacity to delineate widely accepted eff...
| Autores: | , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | Universidad Rey Juan Carlos |
| Repositorio: | BURJC-Digital. Repositorio Institucional de la Universidad Rey Juan Carlos |
| OAI Identifier: | oai:burjcdigital.urjc.es:10115/28339 |
| Acceso en línea: | https://hdl.handle.net/10115/28339 |
| Access Level: | acceso abierto |
| Palabra clave: | Distributed Lag Model Advertising Lagged effects Machine learning Prediction |
| Sumario: | Over the course of several decades, extensive research has been dedicated to Market Response models, often lacking validation in purely predictive tasks and frequently overlooking the adherence of methods to underlying assumptions and conditions, such as the capacity to delineate widely accepted effects of advertising actions. This study introduces an improved method for market response models that aligns with these underlying assumptions. The proposed method is grounded in Distributed Lag Models and distinguishes itself by incorporating regularization in its estimation, employing a cross-validation framework, and implementing hold-out testing. Additionally, it presents an empirical approach to extracting the effects of the model. This methodology facilitates the construction of models in an exploratory and straightforward manner, thereby unlocking the potential to uncover underlying effects and proving suitable for extensive samples and numerous variables. To illustrate its practical application, a real-world data example is provided, accompanied by an unprecedented set of empirical explainability assessments and a high level of predictive capability under circumstances similar to those encountered in corporate decision-making processes. |
|---|