AutocorrelationFitDetector<T, TInput, TOutput>
Models & Types ยท AiDotNet.FitDetectors
A fit detector that analyzes autocorrelation in model residuals to assess model fit.
Inheritance: FitDetectorBase<T, TInput, TOutput> โ AutocorrelationFitDetector<T, TInput, TOutput>
For Beginners
Autocorrelation refers to the correlation of a time series with its own past values. In the context of model fitting, autocorrelation in residuals (prediction errors) can indicate that the model is missing important patterns in the data.
How It Works
This detector uses the Durbin-Watson statistic to measure autocorrelation in model residuals. The Durbin-Watson statistic ranges from 0 to 4, with a value of 2 indicating no autocorrelation, values less than 2 indicating positive autocorrelation, and values greater than 2 indicating negative autocorrelation.
Constructors
| Constructor | Summary |
|---|---|
AutocorrelationFitDetector(AutocorrelationFitDetectorOptions options) |
Initializes a new instance of the AutocorrelationFitDetector class. |
Methods
| Method | Summary |
|---|---|
FitDetectorResult<T> DetectFit(ModelEvaluationData<T, TInput, TOutput> evaluationData) |
Detects the fit type of a model based on autocorrelation in residuals. |