A multiscale model for multivariate time series forecasting
Abstract Transformer based models for time-series forecasting have shown promising performance and during the past few years different Transformer variants have been proposed in time-series forecasting domain. However, most of the existing methods, mainly represent the time-series from a single scal...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-024-82417-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841544705013186560 |
|---|---|
| author | Vahid Naghashi Mounir Boukadoum Abdoulaye Banire Diallo |
| author_facet | Vahid Naghashi Mounir Boukadoum Abdoulaye Banire Diallo |
| author_sort | Vahid Naghashi |
| collection | DOAJ |
| description | Abstract Transformer based models for time-series forecasting have shown promising performance and during the past few years different Transformer variants have been proposed in time-series forecasting domain. However, most of the existing methods, mainly represent the time-series from a single scale, making it challenging to capture various time granularities or ignore inter-series correlations between the series which might lead to inaccurate forecasts. In this paper, we address the above mentioned shortcomings and propose a Transformer based model which integrates multi-scale patch-wise temporal modeling and channel-wise representation. In the multi-scale temporal part, the input time-series is divided into patches of different resolutions to capture temporal correlations associated with various scales. The channel-wise encoder which comes after the temporal encoder, models the relations among the input series to capture the intricate interactions between them. In our framework, we further design a multi-step linear decoder to generate the final predictions for the purpose of reducing over-fitting and noise effects. Extensive experiments on seven real world datasets indicate that our model (MultiPatchFormer) achieves state-of-the-art results by surpassing other current baseline models in terms of error metrics and shows stronger generalizability. |
| format | Article |
| id | doaj-art-dd2723f8fefb46fa942d9a99f6352108 |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-dd2723f8fefb46fa942d9a99f63521082025-01-12T12:22:27ZengNature PortfolioScientific Reports2045-23222025-01-0115111410.1038/s41598-024-82417-4A multiscale model for multivariate time series forecastingVahid Naghashi0Mounir Boukadoum1Abdoulaye Banire Diallo2Computer Science, Université du Québec à MontréalComputer Science, Université du Québec à MontréalComputer Science, Université du Québec à MontréalAbstract Transformer based models for time-series forecasting have shown promising performance and during the past few years different Transformer variants have been proposed in time-series forecasting domain. However, most of the existing methods, mainly represent the time-series from a single scale, making it challenging to capture various time granularities or ignore inter-series correlations between the series which might lead to inaccurate forecasts. In this paper, we address the above mentioned shortcomings and propose a Transformer based model which integrates multi-scale patch-wise temporal modeling and channel-wise representation. In the multi-scale temporal part, the input time-series is divided into patches of different resolutions to capture temporal correlations associated with various scales. The channel-wise encoder which comes after the temporal encoder, models the relations among the input series to capture the intricate interactions between them. In our framework, we further design a multi-step linear decoder to generate the final predictions for the purpose of reducing over-fitting and noise effects. Extensive experiments on seven real world datasets indicate that our model (MultiPatchFormer) achieves state-of-the-art results by surpassing other current baseline models in terms of error metrics and shows stronger generalizability.https://doi.org/10.1038/s41598-024-82417-4 |
| spellingShingle | Vahid Naghashi Mounir Boukadoum Abdoulaye Banire Diallo A multiscale model for multivariate time series forecasting Scientific Reports |
| title | A multiscale model for multivariate time series forecasting |
| title_full | A multiscale model for multivariate time series forecasting |
| title_fullStr | A multiscale model for multivariate time series forecasting |
| title_full_unstemmed | A multiscale model for multivariate time series forecasting |
| title_short | A multiscale model for multivariate time series forecasting |
| title_sort | multiscale model for multivariate time series forecasting |
| url | https://doi.org/10.1038/s41598-024-82417-4 |
| work_keys_str_mv | AT vahidnaghashi amultiscalemodelformultivariatetimeseriesforecasting AT mounirboukadoum amultiscalemodelformultivariatetimeseriesforecasting AT abdoulayebanirediallo amultiscalemodelformultivariatetimeseriesforecasting AT vahidnaghashi multiscalemodelformultivariatetimeseriesforecasting AT mounirboukadoum multiscalemodelformultivariatetimeseriesforecasting AT abdoulayebanirediallo multiscalemodelformultivariatetimeseriesforecasting |