def __init__(self,
Cg=None,
Cn=None,
G=None,
R=None,
rank=1,
Cr=None,
S_R=None,
U_R=None):
"""
Args:
Cg: Limix covariance matrix for Cg (dimension dim_c)
Cn: Limix covariance matrix for Cn (dimension dim_c)
G: [dim_r, rank_r] numpy covariance matrix for G
R: [dim_r, dim_r] numpy semidemidefinite covariance matrix for R
rank: rank of column low-rank covariance (default = 1)
Cr: Limix covariance matrix for Cr (optional).
If not specified, a low-rank covariance matrix is considered
S_R: N vector of eigenvalues of R
U_R: [N, N] eigenvector matrix of R
"""
Covariance.__init__(self)
self._Cr_act = True
self._Cg_act = True
self._Cn_act = True
self.setColCovars(Cg=Cg, Cn=Cn, rank=rank, Cr=None)
self.setR(R=R, S_R=S_R, U_R=U_R)
self.G = G
self.dim = self.dim_c * self.dim_r
self._calcNumberParams()
self._use_to_predict = False
def __init__(self, dim, rank = 1):
"""
Args:
dim: dimension of the low-rank covariance
rank: rank of the low-rank covariance
"""
Covariance.__init__(self)
self.initialize(dim, rank)
def __init__(self, dim, rank=1):
"""
Args:
dim: dimension of the low-rank covariance
rank: rank of the low-rank covariance
"""
Covariance.__init__(self)
self.initialize(dim, rank)
def __init__(self, kinship, design, kinship_cm = None, kinship_cross = None, jitter = 1e-4):
if kinship_cm is None: kinship_cm = kinship
if kinship_cross is None: kinship_cross = kinship
ff_dim = 2
self.covff = FreeFormCov(ff_dim, jitter = 1e-4)
self._K = kinship
self._ZK = sp.dot(design, kinship_cross.T)
self._KZ = sp.dot(kinship_cross, design.T)
self._ZKZ = sp.dot(design, sp.dot(kinship_cm, design.T))
Covariance.__init__(self, kinship.shape[0])
def __init__(self, *covars):
"""
Args:
covars: covariances to be considered in the sum
"""
Covariance.__init__(self)
self.dim = None
self.covars = []
for covar in covars:
self.addCovariance(covar)
covar.register(self.clear_all)
def __init__(self, C, R, Iok=None):
"""
Args:
C: column LIMIX covariance
R: row numpy covariance matrix
"""
Covariance.__init__(self)
self._C = C
self._R = R
self.dim = C.dim * R.shape[0]
self.Iok = Iok
C.register(self.clear_all)
def __init__(self, C, R, Iok=None):
"""
Args:
C: column LIMIX covariance
R: row numpy covariance matrix
"""
Covariance.__init__(self)
self._C = C
self._R = R
self.dim = C.dim * R.shape[0]
self.Iok = Iok
C.register(self.clear_all)
def __init__(self, dim, jitter=1e-4):
"""
Args:
dim: dimension of the free-form covariance
jitter: extent of diagonal offset which is added for numerical stability
(default value: 1e-4)
"""
Covariance.__init__(self, dim)
self._K_act = True
self._calcNumberParams()
self.dim = dim
self.params = sp.zeros(self.n_params)
self.idx_r, self.idx_c = sp.tril_indices(self.dim)
self.set_jitter(jitter)
def __init__(self, Cn = None, G = None, rank = 1):
"""
Args:
Cn: Limix covariance matrix for Cn (dimension dim_c)
G: [dim_r, rank_r] numpy covariance matrix for G
rank: rank of column low-rank covariance (default = 1)
"""
Covariance.__init__(self)
self._Cr_act = True
self._Cn_act = True
self.setColCovars(Cn, rank = rank)
self.G = G
self.dim = self.dim_c * self.dim_r
self._use_to_predict = False
def __init__(self, Cn=None, G=None, rank=1):
"""
Args:
Cn: Limix covariance matrix for Cn (dimension dim_c)
G: [dim_r, rank_r] numpy covariance matrix for G
rank: rank of column low-rank covariance (default = 1)
"""
Covariance.__init__(self)
self._Cr_act = True
self._Cn_act = True
self.setColCovars(Cn, rank=rank)
self.G = G
self.dim = self.dim_c * self.dim_r
self._use_to_predict = False
def __init__(self,
kinship,
design,
kinship_cm=None,
kinship_cross=None,
jitter=1e-4):
if kinship_cm is None: kinship_cm = kinship
if kinship_cross is None: kinship_cross = kinship
ff_dim = 2
self.covff = FreeFormCov(ff_dim, jitter=1e-4)
self._K = kinship
self._ZK = sp.dot(design, kinship_cross.T)
self._KZ = sp.dot(kinship_cross, design.T)
self._ZKZ = sp.dot(design, sp.dot(kinship_cm, design.T))
Covariance.__init__(self, kinship.shape[0])
def __init__(self, Cg=None, Cn=None, R=None, S_R=None, U_R=None):
"""
Args:
Cg: Limix covariance matrix for Cg (dimension dim_c)
Cn: Limix covariance matrix for Cn (dimension dim_c)
R: [dim_r, dim_r] numpy semidemidefinite covariance matrix for R
In alternative to R, S_R and U_R can be specified.
S_R: N vector of eigenvalues of R
U_R: [N, N] eigenvector matrix of R
"""
Covariance.__init__(self)
self._Cg_act = True
self._Cn_act = True
self.setColCovars(Cg, Cn)
self.setR(R=R, S_R=S_R, U_R=U_R)
self.dim = self.dim_c * self.dim_r
self._use_to_predict = False
def __init__(self, Cg = None, Cn = None, R = None, S_R = None, U_R = None):
"""
Args:
Cg: Limix covariance matrix for Cg (dimension dim_c)
Cn: Limix covariance matrix for Cn (dimension dim_c)
R: [dim_r, dim_r] numpy semidemidefinite covariance matrix for R
In alternative to R, S_R and U_R can be specified.
S_R: N vector of eigenvalues of R
U_R: [N, N] eigenvector matrix of R
"""
Covariance.__init__(self)
self._Cg_act = True
self._Cn_act = True
self.setColCovars(Cg, Cn)
self.setR(R=R, S_R=S_R, U_R=U_R)
self.dim = self.dim_c * self.dim_r
self._use_to_predict = False
def __init__(self, K0, Kcross0=None):
"""
Args:
K0: semi-definite positive matrix that defines the fixed-form covariance
Kcross0: cross covariance between training and test samples
(used only for out-of-sample predictions)
"""
Covariance.__init__(self)
self._scale_act = True
self.K0 = assert_make_float_array(K0, "K0")
assert_finite_array(self.K0)
if Kcross0 is not None:
Kcross0 = assert_make_float_array(Kcross0, "Kcross0")
assert_finite_array(Kcross0)
self.Kcross0 = Kcross0
self.params = np.zeros(1)
def __init__(self, X, Xstar=None):
"""
X: [dim, 1] input matrix
Xstar: [dim_star, 1] out-of-sample input matrix
"""
Covariance.__init__(self)
self._scale_act = True
self._length_act = True
X = assert_make_float_array(X, "X")
assert_finite_array(X)
self.X = X
if Xstar is not None:
Xstar = assert_make_float_array(Xstar, "Xstar")
assert_finite_array(Xstar)
self.Xstar = Xstar
self.params = np.zeros(2)
def __init__(self, X, Xstar=None):
"""
X: [dim, 1] input matrix
Xstar: [dim_star, 1] out-of-sample input matrix
"""
Covariance.__init__(self)
self._scale_act = True
self._length_act = True
X = assert_make_float_array(X, "X")
assert_finite_array(X)
self.X = X
if Xstar is not None:
Xstar = assert_make_float_array(Xstar, "Xstar")
assert_finite_array(Xstar)
self.Xstar = Xstar
self.params = np.zeros(2)
def __init__(self, Cg=None, Cn=None, G=None, R=None, rank=1, Cr=None, S_R=None, U_R=None):
"""
Args:
Cg: Limix covariance matrix for Cg (dimension dim_c)
Cn: Limix covariance matrix for Cn (dimension dim_c)
G: [dim_r, rank_r] numpy covariance matrix for G
R: [dim_r, dim_r] numpy semidemidefinite covariance matrix for R
rank: rank of column low-rank covariance (default = 1)
Cr: Limix covariance matrix for Cr (optional).
If not specified, a low-rank covariance matrix is considered
S_R: N vector of eigenvalues of R
U_R: [N, N] eigenvector matrix of R
"""
Covariance.__init__(self)
self._Cr_act = True
self._Cg_act = True
self._Cn_act = True
self.setColCovars(Cg=Cg, Cn=Cn, rank=rank, Cr=None)
self.setR(R=R, S_R=S_R, U_R=U_R)
self.G = G
self.dim = self.dim_c * self.dim_r
self._calcNumberParams()
self._use_to_predict = False
def __init__(self):
Covariance.__init__(self)
self.covars = []
def __init__(self, gp):
Covariance.__init__(self)
self.gp = gp
gp.register(self.clear_all)
self.dim = gp.mean.b.shape[0]
