An F# library for model-fitting model-selection (MFMS) of time-series models to observational data. Included is expressive modelling system for authoring models, which includes full F# units of measure support.
Bristlecone contains specialist functionality for fitting (long-term) ecological models. The library was developed for use in dendroecology and palaeoecology, but can be used for other time-series modelling problems.
| Package | Stable | Prerelease |
|---|---|---|
| Bristlecone |  |
 |
| Bristlecone.Dendro |  |
 |
To get started with Bristlecone, see the documentation website with tutorials, examples and API specification.
Some examples are provided on the docs page, with optional downloadable worked notebooks.
A suite of benchmarks - covering optimisation methods and ecological time-series models - are included in the test/Bristlecone.Benchmark project. The most recent results of the benchmark tests are here.
Bristlecone requires the .NET 10 SDK, and uses FAKE for building, testing, documentation, and publishing.
Simple building and testing can be done using dotnet build or dotnet run test/Bristlecone.Tests/Bristlecone.Tests.fsproj respectively in the root directory. Docs can be tested by first calling dotnet tool restore then
dotnet fsdocs watch --eval, which serves them on a Suave server.
The full build process can be run using either dotnet fsi build.fsx, which uses FAKE to run all build steps,
tests, and compile docs.
Contributions, issues, and pull-requests are welcome.
Logo based on Free Vector Art by www.vecteezy.com
