Class BilingualChiSquareComputer3
Inheritance
System.Object
BilingualChiSquareComputer3
Namespace: Sdl.Core.FineGrainedAlignment
Assembly: Sdl.Core.FineGrainedAlignment.dll
Syntax
public class BilingualChiSquareComputer3 : ObjectConstructors
BilingualChiSquareComputer3(DataLocation2, IEnumerable<Pair<IntSegment>>)
Declaration
public BilingualChiSquareComputer3(DataLocation2 location, IEnumerable<Pair<IntSegment>> segmentReaders)Parameters
| Type | Name | Description |
|---|---|---|
| Sdl.LanguagePlatform.Stat.DataLocation2 | location | |
| System.Collections.Generic.IEnumerable<Pair<Sdl.LanguagePlatform.Stat.IntSegment>> | segmentReaders |
Fields
Progress
Declaration
public EventHandler<TranslationModelProgressEventArgs> ProgressField Value
| Type | Description |
|---|---|
| System.EventHandler<TranslationModelProgressEventArgs> |
Methods
Compute()
Computes chi-square scores for all token pairs in the corpus, using default parameters.
Declaration
public SparseMatrix<double> Compute()Returns
| Type | Description |
|---|---|
| Sdl.LanguagePlatform.Stat.SparseMatrix<System.Double> | A matrix with the scores. Empty cells are below the min score |
Compute(Double, Int32, Int32)
Computes chi-square scores for all token pairs of the corpus. Can take very long (several hours) to compute on larger TMs (200.000 TUs and up).
Declaration
public SparseMatrix<double> Compute(double minScore, int minFreq, int minCooc)Parameters
| Type | Name | Description |
|---|---|---|
| System.Double | minScore | The minimum chi-square score a token pair must have to be included in the result. 5% ~= minscore = 3.841 (should be the minimum) |
| System.Int32 | minFreq | minimum occurrence (frequency) of a token to be considered |
| System.Int32 | minCooc | minimum co-occurrence of a token pair to be considered |
Returns
| Type | Description |
|---|---|
| Sdl.LanguagePlatform.Stat.SparseMatrix<System.Double> | A matrix with the scores. Empty cells are below the min score |