ChiSquareDetector<T>
Models & Types · AiDotNet.AnomalyDetection.Statistical
Detects anomalies using the Chi-Square test for multivariate data.
Inheritance: ModelBase<T, Matrix<T>, Vector<T>> → AnomalyDetectorBase<T> → ChiSquareDetector<T>
Implements: IGeneratedParameterRegistrar<T>
For Beginners
The Chi-Square (Mahalanobis distance) detector identifies outliers by measuring how far a point is from the center of the data, accounting for correlations between features. Points far from the center in Mahalanobis distance are anomalies.
How It Works
The algorithm computes: D^2 = (x - mean)’ * Cov^(-1) * (x - mean) This squared Mahalanobis distance follows a Chi-Square distribution with p degrees of freedom (where p = number of features) under the assumption of multivariate normality.
When to use:
- Multivariate data (multiple correlated features)
- Data is approximately multivariate normal
- You want to account for correlations between features
Industry Standard Defaults:
- Alpha (significance level): 0.05 (5%)
Constructors
| Constructor | Summary |
|---|---|
ChiSquareDetector(double alpha, double contamination, int randomSeed) |
Creates a new Chi-Square anomaly detector. |
Properties
| Property | Summary |
|---|---|
T ChiSquareCritical { get; } |
Gets the Chi-Square critical value based on the fitted degrees of freedom (number of features) and alpha. |
double Alpha { get; } |
Gets the significance level (alpha) for the test. |
Methods
| Method | Summary |
|---|---|
Vector<T> Predict(Matrix<T> X) |
|
Vector<T> ScoreAnomalies(Matrix<T> X) |
|
void Fit(Matrix<T> X) |