Exponentially Weighted Information Criteria for Selecting Among Forecasting Models
Information criteria (IC) are often used to decide between forecasting models. Commonly used criteria include Akaike's IC and Schwarz's Bayesian IC. They involve the sum of two terms: the model's log likelihood and a penalty for the number of model parameters. The likelihood is calcul...
| 主要作者: | |
|---|---|
| 格式: | Journal article |
| 出版: |
2008
|
| _version_ | 1826287361643773952 |
|---|---|
| author | Taylor, J |
| author_facet | Taylor, J |
| author_sort | Taylor, J |
| collection | OXFORD |
| description | Information criteria (IC) are often used to decide between forecasting models. Commonly used criteria include Akaike's IC and Schwarz's Bayesian IC. They involve the sum of two terms: the model's log likelihood and a penalty for the number of model parameters. The likelihood is calculated with equal weight being given to all observations. We propose that greater weight should be put on more recent observations in order to reflect more recent accuracy. This seems particularly pertinent when selecting among exponential smoothing methods, as they are based on an exponential weighting principle. In this paper, we use exponential weighting within the calculation of the log likelihood for the IC. Our empirical analysis uses supermarket sales and call centre arrivals data. The results show that basing model selection on the new exponentially weighted IC can outperform individual models and selection based on the standard IC. |
| first_indexed | 2024-03-07T01:57:31Z |
| format | Journal article |
| id | oxford-uuid:9c3a4254-d654-4bfc-87b9-420e5a36e1cf |
| institution | University of Oxford |
| last_indexed | 2024-03-07T01:57:31Z |
| publishDate | 2008 |
| record_format | dspace |
| spelling | oxford-uuid:9c3a4254-d654-4bfc-87b9-420e5a36e1cf2022-03-27T00:34:32ZExponentially Weighted Information Criteria for Selecting Among Forecasting ModelsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:9c3a4254-d654-4bfc-87b9-420e5a36e1cfSaïd Business School - Eureka2008Taylor, JInformation criteria (IC) are often used to decide between forecasting models. Commonly used criteria include Akaike's IC and Schwarz's Bayesian IC. They involve the sum of two terms: the model's log likelihood and a penalty for the number of model parameters. The likelihood is calculated with equal weight being given to all observations. We propose that greater weight should be put on more recent observations in order to reflect more recent accuracy. This seems particularly pertinent when selecting among exponential smoothing methods, as they are based on an exponential weighting principle. In this paper, we use exponential weighting within the calculation of the log likelihood for the IC. Our empirical analysis uses supermarket sales and call centre arrivals data. The results show that basing model selection on the new exponentially weighted IC can outperform individual models and selection based on the standard IC. |
| spellingShingle | Taylor, J Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title | Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title_full | Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title_fullStr | Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title_full_unstemmed | Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title_short | Exponentially Weighted Information Criteria for Selecting Among Forecasting Models |
| title_sort | exponentially weighted information criteria for selecting among forecasting models |
| work_keys_str_mv | AT taylorj exponentiallyweightedinformationcriteriaforselectingamongforecastingmodels |