SeminormalFitter
Class for fitting a 'seminormal' distribution to a point cloud
This is not strictly speaking a probability distribution, only the first quadrant of the result of fitting a normal distribution to the data + its mirror image wrt the origin.
[class] SeminormalFitter
comments:
class for fitting a 'seminormal' distribution to a point cloud
this is not strictly speaking a probability distribution,
only the first quadrant of the result of fitting a normal distribution
to the data + its mirror image wrt the origin.
parameters
- cov: multidim covariance matrix of normal distribution
- mvn: scipy.stats multivariate_normal object built from the cov
⤷ __init__
full signature:
def __init__(self)
comments:
empty constructor
⤷ fit
full signature:
def fit(self,points)
comments:
make the fit
input arguments:
- points: a np array of shape (npoints,ndims)
full signature:
def pdf(self,points)
comments:
get pdf at points
input arguments:
- points: a np array of shape (npoints,ndims)
⤷ save
full signature:
def save(self,path)
comments:
save the covariance matrix as a .npy file specified by path
⤷ load
full signature:
def load(self,path)
comments:
load a covariance matrix from a .npy file specified by path and build the fit from it