teetool.model.Model Class Reference

Model class provides the interface to the probabilistic modelling of trajectories. More...

Inheritance diagram for teetool.model.Model:

Public Member Functions
def	__init__ (self, cluster_data, settings)
	Initialise a model. More...
def	clear (self)
def	getMean (self)
def	getSamples (self, nsamples)
def	getKS (self, cluster_data, sigma_arr, nsamples=100)
def	isInside_grid (self, sdwidth, xx, yy, zz=None)
def	isInside_pnts (self, P, sdwidth=1, nellipse=50)
def	evalLogLikelihood (self, xx, yy, zz=None)
def	getOutline (self, sdwidth=1)

Private Member Functions
def	_norm2real (self, mu_y, sig_y, outline)
def	_clusterdata2points (self, cluster_data)
def	_getEllipse (self, c, A, sdwidth=1, nellipse=10)
def	_getSample (self, c, A, nsamples=1, std=1)
def	_getCoordsEllipse (self, nellipse=20, sdwidth=5)
def	_eval_logp (self, Y_pos)
def	_grid2points (self, xx, yy, zz=None)
def	_points2grid (self, s, Y_idx)
def	_get_point_cloud (self, sdwidth=1, nellipse=50)

Private Attributes
	_ndim
	dimension of cluster_data More...
	_mu_y
	mean vector [ngaus*ndim x 1] More...
	_sig_y
	covariance matrix [ngaus*ndim x ngaus*ndim] More...
	_cc
	mean vector [ndim x 1] in ngaus cells More...
	_cA
	covariance matrix [ndim x ndim] in ngaus cells More...
	_list_tube
	previously calculated values confidence region More...
	_list_logp
	previously calculated values log-likelihood More...

Detailed Description

Model class provides the interface to the probabilistic modelling of trajectories.

These trajectories are a single ensemble

Constructor & Destructor Documentation

__init__()

def teetool.model.Model.__init__	(		self,
			cluster_data,
			settings
	)

Initialise a model.

Parameters

self	object pointer
cluster_data	list of (x, Y) trajectory data
settings	a dictionary with "model_type" = resampling, ML, or EM "ngaus": number of Gaussians to create for output, REQUIRED for ML and EM, "basis_type" = gaussian, bernstein, "nbasis": number of basis functions

Member Function Documentation

_clusterdata2points()

def teetool.model.Model._clusterdata2points (   self,   cluster_data  )

private

returns the points Y [npoints x ndim]

_eval_logp()

def teetool.model.Model._eval_logp (   self,   Y_pos  )

private

evaluates on a grid, aiming at the desired number of points

_get_point_cloud()

def teetool.model.Model._get_point_cloud (   self,   sdwidth = 1,   nellipse = 50  )

private

returns a list with point clouds, representing the transition between Gaussians

input paramters:
    - none

_getCoordsEllipse()

def teetool.model.Model._getCoordsEllipse (   self,   nellipse = 20,   sdwidth = 5  )

private

returns an array of xy(z) coordinates

nellipse is number of points in ellipsoid and sdwidth is the variance

_getEllipse()

def teetool.model.Model._getEllipse (   self,   c,   A,   sdwidth = 1,   nellipse = 10  )

private

evaluates the ellipse

_getSample()

def teetool.model.Model._getSample (   self,   c,   A,   nsamples = 1,   std = 1  )

private

returns nsamples samples of the given Gaussian

_grid2points()

def teetool.model.Model._grid2points (   self,   xx,   yy,   zz = None  )

private

returns an Y matrix for a given grid

xx, yy, (zz) are mgrid

_norm2real()

def teetool.model.Model._norm2real (   self,   mu_y,   sig_y,   outline  )

private

returns the normalised values back to the original

_points2grid()

def teetool.model.Model._points2grid (   self,   s,   Y_idx  )

private

converts points to a matrix

s is values np.array and Y_idx is position np.array

clear()

def teetool.model.Model.clear

(

self

)

clears memory usage

evalLogLikelihood()

def teetool.model.Model.evalLogLikelihood	(		self,
			xx,
			yy,
			zz = None
	)

evaluates values in this grid [2d/3d] and returns values

example grid:
xx, yy, zz = np.mgrid[-60:60:20j, -10:240:20j, -60:60:20j]

getKS()

def teetool.model.Model.getKS	(		self,
			cluster_data,
			sigma_arr,
			nsamples = 100
	)

calculated the ks value

input:
    cluster_data    - data to evaluate against
    nsamples        - number of samples to use
    sigma_arr       - array which standard deviation to evalute

getMean()

def teetool.model.Model.getMean

(

self

)

returns the average trajectory [x, y, (z)]

getOutline()

def teetool.model.Model.getOutline	(		self,
			sdwidth = 1
	)

returns the outline [xmin, xmax, ymin, ymax, zmin, zmax]

input parameters:
    - sdwidth

getSamples()

def teetool.model.Model.getSamples	(		self,
			nsamples
	)

return nsamples of the model

isInside_grid()

def teetool.model.Model.isInside_grid	(		self,
			sdwidth,
			xx,
			yy,
			zz = None
	)

evaluate if points are inside a grid

Input parameters:
    - sdwidth
    - xx
    - yy
    - zz (when 3d)

isInside_pnts()

def teetool.model.Model.isInside_pnts	(		self,
			P,
			sdwidth = 1,
			nellipse = 50
	)

tests if points P NxD 'points' x 'dimensions' are inside the tube

Member Data Documentation

_cA

teetool.model.Model._cA

private

covariance matrix [ndim x ndim] in ngaus cells

_cc

teetool.model.Model._cc

private

mean vector [ndim x 1] in ngaus cells

_list_logp

teetool.model.Model._list_logp

private

previously calculated values log-likelihood

_list_tube

teetool.model.Model._list_tube

private

previously calculated values confidence region

_mu_y

teetool.model.Model._mu_y

private

mean vector [ngaus*ndim x 1]

_ndim

teetool.model.Model._ndim

private

dimension of cluster_data

_sig_y

teetool.model.Model._sig_y

private

covariance matrix [ngaus*ndim x ngaus*ndim]

---

The documentation for this class was generated from the following file:

• teetool/model.py