Learning and filtering via simulation: smoothly jittered particle filters.
A key ingredient of many particle filters is the use of the sampling importance resampling algorithm (SIR), which transforms a sample of weighted draws from a prior distribution into equally weighted draws from a posterior distribution. We give a novel analysis of the SIR algorithm and analyse the...
| Autores principales: | , |
|---|---|
| Formato: | Working paper |
| Lenguaje: | English |
| Publicado: |
Department of Economics (University of Oxford)
2009
|
| _version_ | 1826270614736863232 |
|---|---|
| author | Flury, T Shephard, N |
| author_facet | Flury, T Shephard, N |
| author_sort | Flury, T |
| collection | OXFORD |
| description | A key ingredient of many particle filters is the use of the sampling importance resampling algorithm (SIR), which transforms a sample of weighted draws from a prior distribution into equally weighted draws from a posterior distribution. We give a novel analysis of the SIR algorithm and analyse the jittered generalisation of SIR, showing that existing implementations of jittering lead to marked inferior behaviour over the base SIR algorithm. We show how jittering can be designed to improve the performance of the SIR algorithm. We illustrate its performance in practice in the context of three filtering problems. |
| first_indexed | 2024-03-06T21:43:37Z |
| format | Working paper |
| id | oxford-uuid:48caed1d-b7a1-47e8-a2b1-5944d9b0741c |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T21:43:37Z |
| publishDate | 2009 |
| publisher | Department of Economics (University of Oxford) |
| record_format | dspace |
| spelling | oxford-uuid:48caed1d-b7a1-47e8-a2b1-5944d9b0741c2022-03-26T15:27:47ZLearning and filtering via simulation: smoothly jittered particle filters.Working paperhttp://purl.org/coar/resource_type/c_8042uuid:48caed1d-b7a1-47e8-a2b1-5944d9b0741cEnglishDepartment of Economics - ePrintsDepartment of Economics (University of Oxford)2009Flury, TShephard, NA key ingredient of many particle filters is the use of the sampling importance resampling algorithm (SIR), which transforms a sample of weighted draws from a prior distribution into equally weighted draws from a posterior distribution. We give a novel analysis of the SIR algorithm and analyse the jittered generalisation of SIR, showing that existing implementations of jittering lead to marked inferior behaviour over the base SIR algorithm. We show how jittering can be designed to improve the performance of the SIR algorithm. We illustrate its performance in practice in the context of three filtering problems. |
| spellingShingle | Flury, T Shephard, N Learning and filtering via simulation: smoothly jittered particle filters. |
| title | Learning and filtering via simulation: smoothly jittered particle filters. |
| title_full | Learning and filtering via simulation: smoothly jittered particle filters. |
| title_fullStr | Learning and filtering via simulation: smoothly jittered particle filters. |
| title_full_unstemmed | Learning and filtering via simulation: smoothly jittered particle filters. |
| title_short | Learning and filtering via simulation: smoothly jittered particle filters. |
| title_sort | learning and filtering via simulation smoothly jittered particle filters |
| work_keys_str_mv | AT fluryt learningandfilteringviasimulationsmoothlyjitteredparticlefilters AT shephardn learningandfilteringviasimulationsmoothlyjitteredparticlefilters |