Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation.
Calibrating model parameters on heterogeneous data can be challenging and inefficient. This holds especially for likelihood-free methods such as approximate Bayesian computation (ABC), which rely on the comparison of relevant features in simulated and observed data and are popular for otherwise intr...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2023-01-01
|
| Series: | PLoS ONE |
| Online Access: | https://doi.org/10.1371/journal.pone.0285836 |
| _version_ | 1827923558138904576 |
|---|---|
| author | Yannik Schälte Jan Hasenauer |
| author_facet | Yannik Schälte Jan Hasenauer |
| author_sort | Yannik Schälte |
| collection | DOAJ |
| description | Calibrating model parameters on heterogeneous data can be challenging and inefficient. This holds especially for likelihood-free methods such as approximate Bayesian computation (ABC), which rely on the comparison of relevant features in simulated and observed data and are popular for otherwise intractable problems. To address this problem, methods have been developed to scale-normalize data, and to derive informative low-dimensional summary statistics using inverse regression models of parameters on data. However, while approaches only correcting for scale can be inefficient on partly uninformative data, the use of summary statistics can lead to information loss and relies on the accuracy of employed methods. In this work, we first show that the combination of adaptive scale normalization with regression-based summary statistics is advantageous on heterogeneous parameter scales. Second, we present an approach employing regression models not to transform data, but to inform sensitivity weights quantifying data informativeness. Third, we discuss problems for regression models under non-identifiability, and present a solution using target augmentation. We demonstrate improved accuracy and efficiency of the presented approach on various problems, in particular robustness and wide applicability of the sensitivity weights. Our findings demonstrate the potential of the adaptive approach. The developed algorithms have been made available in the open-source Python toolbox pyABC. |
| first_indexed | 2024-03-13T04:59:38Z |
| format | Article |
| id | doaj.art-58e96bf5af50430fbc499e4756a4c8c6 |
| institution | Directory Open Access Journal |
| issn | 1932-6203 |
| language | English |
| last_indexed | 2024-03-13T04:59:38Z |
| publishDate | 2023-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj.art-58e96bf5af50430fbc499e4756a4c8c62023-06-17T05:31:12ZengPublic Library of Science (PLoS)PLoS ONE1932-62032023-01-01185e028583610.1371/journal.pone.0285836Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation.Yannik SchälteJan HasenauerCalibrating model parameters on heterogeneous data can be challenging and inefficient. This holds especially for likelihood-free methods such as approximate Bayesian computation (ABC), which rely on the comparison of relevant features in simulated and observed data and are popular for otherwise intractable problems. To address this problem, methods have been developed to scale-normalize data, and to derive informative low-dimensional summary statistics using inverse regression models of parameters on data. However, while approaches only correcting for scale can be inefficient on partly uninformative data, the use of summary statistics can lead to information loss and relies on the accuracy of employed methods. In this work, we first show that the combination of adaptive scale normalization with regression-based summary statistics is advantageous on heterogeneous parameter scales. Second, we present an approach employing regression models not to transform data, but to inform sensitivity weights quantifying data informativeness. Third, we discuss problems for regression models under non-identifiability, and present a solution using target augmentation. We demonstrate improved accuracy and efficiency of the presented approach on various problems, in particular robustness and wide applicability of the sensitivity weights. Our findings demonstrate the potential of the adaptive approach. The developed algorithms have been made available in the open-source Python toolbox pyABC.https://doi.org/10.1371/journal.pone.0285836 |
| spellingShingle | Yannik Schälte Jan Hasenauer Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. PLoS ONE |
| title | Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. |
| title_full | Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. |
| title_fullStr | Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. |
| title_full_unstemmed | Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. |
| title_short | Informative and adaptive distances and summary statistics in sequential approximate Bayesian computation. |
| title_sort | informative and adaptive distances and summary statistics in sequential approximate bayesian computation |
| url | https://doi.org/10.1371/journal.pone.0285836 |
| work_keys_str_mv | AT yannikschalte informativeandadaptivedistancesandsummarystatisticsinsequentialapproximatebayesiancomputation AT janhasenauer informativeandadaptivedistancesandsummarystatisticsinsequentialapproximatebayesiancomputation |