|
MatrixRectangular | Model::evalCovMatrix (Db *db1, Db *db2=nullptr, int ivar0=-1, int jvar0=-1, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr) |
|
MatrixSquareSymmetric | Model::evalCovMatrixSymmetric (Db *db1, int ivar0=-1, const VectorInt &nbgh1=VectorInt(), const CovCalcMode *mode=nullptr) |
|
MatrixSparse * | Model::evalCovMatrixSparse (Db *db1, Db *db2=nullptr, int ivar0=-1, int jvar0=-1, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr, double eps=EPSILON3) |
|
VectorDouble | Model::evalCovMatrixV (Db *db1, Db *db2=nullptr, int ivar0=-1, int jvar0=-1, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr) |
|
MatrixRectangular | Model::evalCovMatrixOptim (const Db *db1, const Db *db2=nullptr, int ivar0=-1, int jvar0=-1, const VectorInt &nbgh1=VectorInt(), const VectorInt &nbgh2=VectorInt(), const CovCalcMode *mode=nullptr) |
|
MatrixSquareSymmetric | Model::evalCovMatrixSymmetricOptim (const Db *db1, int ivar0=-1, const VectorInt &nbgh1=VectorInt(), const CovCalcMode *mode=nullptr) |
|
These functions are meant to calculate the covariance Matrix between two Dbs or between a Db and itself. They take into account the presence of a possible selection They also account for heterotopy (if Z-variables are defined in the Db(s)
- Parameters
-
db1 | First Db |
db2 | (Optional second Db) |
ivar0 | Rank of the selected variable in the first Db (-1 for all variables) |
jvar0 | Rank of the selected variable in the second Db (-1 for all variables) |
nbgh1 | Vector of indices of active samples in first Db (optional) |
nbgh2 | Vector of indices of active samples in second Db (optional) |
mode | CovCalcMode structure |
- Note
- 'dbin' and 'dbout' cannot be made 'const' as they can be updated
-
due to the presence of 'nostat'
- Returns
- A Matrix either in Dense or Sparse format