https://github.com/hauselin/firdeconvolution
Python class that performs finite impulse response fitting on time series data.
Science Score: 23.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
○codemeta.json file
-
○.zenodo.json file
-
✓DOI references
Found 4 DOI reference(s) in README -
✓Academic publication links
Links to: zenodo.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (8.2%) to scientific vocabulary
Last synced: 5 months ago
·
JSON representation
Repository
Python class that performs finite impulse response fitting on time series data.
Basic Info
- Host: GitHub
- Owner: hauselin
- License: mit
- Default Branch: master
- Homepage: tknapen.github.io/FIRDeconvolution
- Size: 10.4 MB
Statistics
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
- Releases: 0
Fork of tknapen/FIRDeconvolution
Created over 5 years ago
· Last pushed over 8 years ago
https://github.com/hauselin/FIRDeconvolution/blob/master/
# FIRDeconvolution FIRDeconvolution is a python class that performs finite impulse response fitting on time series data, in order to estimate event-related signals. Example use cases are fMRI and pupil size analysis. The package performs the linear least squares analysis using numpy.linalg as a backend, but can switch between different backends, such as statsmodels (which is implemented). For very collinear design matrices ridge regression is implemented through the sklearn RidgeCV function. Bootstrap estimates of error regions are implemented through residual reshuffling. It is possible to add covariates to the events to estimate not just the impulse response function, but also correlation timecourses with secondary variables. Furthermore, one can add the duration each event should have in the designmatrix, for designs in which the durations of the events vary. In neuroscience, the inspection of the event-related signals such as those estimated by FIRDeconvolution is essential for a thorough understanding of one's data. Researchers may overlook essential patterns in their data when blindly running GLM analyses without looking at the impulse response shapes. The test notebook explains how the package can be used for data analysis, by creating toy signals and then using FIRDeconvolution to fit the impulse response functions from the toy data. ## Dependencies numpy, scipy, matplotlib, statsmodels, sklearn TODO - temporal autocorrelation correction [](http://dx.doi.org/10.5281/zenodo.46216)
Owner
- Name: Hause Lin
- Login: hauselin
- Kind: user
- Website: hauselin.com
- Twitter: hauselin
- Repositories: 182
- Profile: https://github.com/hauselin
Researcher at MIT & World Bank