SignificanceAnalysis<T>

Models & Types · AiDotNet.Preprocessing.FeatureSelection.Bioinformatics

Significance Analysis of Microarrays (SAM) for bioinformatics feature selection.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → SignificanceAnalysis<T>

Implements: IFixedCountFeatureSelector<T>, ISupervisedFeatureSelector<T>

For Beginners

When analyzing gene expression data, some genes have very small changes but even smaller variance, making them appear significant by chance. SAM adds a small buffer to prevent these false alarms. It also uses shuffling to estimate how many of our “discoveries” might be false.

How It Works

SAM uses a modified t-statistic that includes a small constant (fudge factor) in the denominator to avoid false positives from genes with very low variance. It estimates false discovery rate (FDR) using permutation testing.

Constructors

Constructor Summary
SignificanceAnalysis(int nFeaturesToSelect, int nPermutations, double fdrThreshold, int? randomState, int[] columnIndices)

Properties

Property Summary
Vector<T> QValues { get; }
Vector<T> SAMScores { get; }
bool SupportsInverseTransform { get; }
double FDRThreshold { get; }
int NFeaturesToSelect { get; }
int NPermutations { get; }
int[] SelectedIndices { get; }

Methods

Method Summary
Matrix<T> FitTransform(Matrix<T> data, Vector<T> target)
bool[] GetSupportMask()
string[] GetFeatureNamesOut(string[] inputFeatureNames)
void Fit(Matrix<T> data, Vector<T> target)