Skip to content

getting-started.md

Latest commit

 

History

History
98 lines (69 loc) · 2.61 KB

getting-started.md

File metadata and controls

98 lines (69 loc) · 2.61 KB
Raw

Getting Started

Back to Index | Next: Models Overview ->

This guide shows the minimum namespaces and workflows needed to use RMC.BestFit.dll from a .NET application.

Installation

Reference the BestFit model library and its Numerics dependency from your application:

<ProjectReference Include="..\RMC.BestFit\RMC.BestFit.csproj" />

BestFit uses Numerics for probability distributions, optimization, MCMC, and matrix operations.

Required Namespaces

using Numerics.Distributions;
using RMC.BestFit;
using RMC.BestFit.Analyses;
using RMC.BestFit.Diagnostics;
using RMC.BestFit.Estimation;
using RMC.BestFit.Models;

RMC.BestFit contains observation and series types such as ExactSeries, while RMC.BestFit.Models contains DataFrame and model classes.

Create Input Data

using RMC.BestFit;
using RMC.BestFit.Models;

var dataFrame = new DataFrame
{
    ExactSeries = new ExactSeries(new[] { 22.0, 27.4, 31.8, 29.1, 35.7 })
};

For censored or uncertain records, use UncertainSeries, IntervalSeries, and ThresholdSeries.

Fit One Distribution With MLE

using Numerics.Distributions;
using RMC.BestFit;
using RMC.BestFit.Estimation;
using RMC.BestFit.Models;

var dataFrame = new DataFrame
{
    ExactSeries = new ExactSeries(new[] { 22.0, 27.4, 31.8, 29.1, 35.7 })
};

var model = new UnivariateDistribution(dataFrame, UnivariateDistributionType.Normal);
var mle = new MaximumLikelihood(model, OptimizationMethod.DifferentialEvolution);

if (mle.Estimate())
{
    model.SetParameterValues(mle.BestParameterSet.Values);
    Console.WriteLine($"Log-likelihood: {mle.MaximumLogLikelihood:F3}");
}

Run Bayesian Analysis

using Numerics.Distributions;
using RMC.BestFit;
using RMC.BestFit.Analyses;
using RMC.BestFit.Models;

var dataFrame = new DataFrame
{
    ExactSeries = new ExactSeries(new[] { 22.0, 27.4, 31.8, 29.1, 35.7 })
};

var model = new UnivariateDistribution(dataFrame, UnivariateDistributionType.LogPearsonTypeIII);
var analysis = new UnivariateAnalysis(model);

analysis.BayesianAnalysis.Iterations = 3000;
analysis.BayesianAnalysis.WarmupIterations = 1500;

await analysis.RunAsync();

Console.WriteLine($"DIC: {analysis.BayesianAnalysis.DIC:F2}");
Console.WriteLine($"WAIC: {analysis.BayesianAnalysis.WAIC:F2}");

Next Steps

Read Models Overview for the model contract, then Input Data Frame for flood-frequency data types.

Back to Index | Next: Models Overview ->