Estimate discontinuous timeseries from continuous covariates.
Project description
discontinuum
[!WARNING]
Experimental.
Overview
discontinuum
is a middleware for developing Gaussian process (GP) timeseries models.
Why might we want a middleware?
GP's are an elegant way to model timeseries with uncertainty.
In many cases, we can represent a complex timeseries as a GP with only a few lines of math.
However, fitting GP's is numerically intense, $\mathcal{O}(n^3)$ complexity.
There are several optimizations that take advantage of simplifying assumptions, different algorithms, or GPUs,
but each has different tradeoffs.
Ideally, we could write the mathematical model once, then run it on whichever "engine" is best suited for a particular problem.
With every model comes a lot of standard utility functions,
and the goal of discontinuum
is to package these different model applications, engines, and utilities into a single ecosystem.
Installation
pip install discontinuum
Models
Only one for now.
loadset-gp
LOAD ESTimator (LOADEST) is a software program for estimating river constituent timeseries using surrogate variables (covariates).
For example, estimating nitrate concentration based on date and streamflow.
However, LOADEST has several serious limitations---it's essentially a linear regression---
and it has been all but replaced by the more flexible Weighted Regression on Time Discharge and Season (WRTDS),
which allows the relation between target and covariate to vary through time.
loadest-gp
takes the WRTDS idea and reimplements it as a GP.
from loadest_gp import LoadestGP()
model = LoadestGP()
model.fit(target, covariates)
model.plot(covariates)
Engines
Currently, the only engine is pymc
's marginal likelihood implementation.
Roadmap
mindmap
root((discontinuum))
data providers
USGS
EPA
etc
engines
PyMC
Tensorflow
PyTorch
utilities
pre-processing
post-processing
plotting
models
loadest-gp
your own
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for discontinuum-0.4-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | c9b46fc6c472b7cc1886ead8c3cbe899646020e3b2e00a931b2f903001cfe35a |
|
MD5 | 010fbbc0449b189c3bd53fa8bdadee43 |
|
BLAKE2b-256 | 699099c9b80c28d84e9ca3e7adb25117488f79f6994b6e227c0b0646b47a9342 |