API Reference¶
This is the class and function reference for mouse_connectivity_models. Please refer to the full user guide for further details.
mcmodels.core
: Core data related classes and utility functions¶
The mcmodels.core
module implements objects useful in data manipulation.
User Guide See the Working with data from the AllenSDK section for further details.
core.VoxelData (cache[, …]) |
Container class for voxel-scale grid data. |
core.RegionalData (cache[, …]) |
Container class for regionalized voxel-scale grid data. |
core.Experiment ([injection_density, …]) |
Class containing the data from an anterograde injection |
core.Mask (reference_space[, structure_ids, …]) |
Object for masking the grid data from allensdk. |
core.VoxelModelCache ([resolution, cache, …]) |
Cache class extending MouseConnectivityCache to cache voxel model data. |
core.VoxelModelApi ([base_uri]) |
HTTP Client extending MouseConnectivityApi to download model data. |
Utitility fucntions¶
Module containing utility functions for the mcmodels.core
module.
core.utils.compute_centroid (injection_density) |
Computes centroid in index coordinates. |
core.utils.get_injection_hemisphere_id (…) |
Gets injection hemisphere based on injection density. |
mcmodels.regressors
: Classes for Performing Regressions¶
The mcmodels.regressors
module implements scikit-learn style estimators
for solving various regression problems. It implements the NadarayaWatson
regressor, and both regularized (L2) and non-regularized non-negative least
squares linear models.
mcmodels.regressors.nonnegative_linear
: Nonnegative Linear Least squares¶
The mcmodels.regressors.nonnegative_linear
module implements linear
models subject to the nonnegativity constraint. It includes Nonnegative linear
regression and experimental modules implementing Nonnegative linear regression
with L2 (Ridge) regularization.
Note
- If one wishes to perform non-negative Lasso regression, see sklearn.linear_model.Lasso or sklearn.linear_model.lasso_path and pass the parameters fit_intercept=False, positive=True
- If one wishes to perform non-negative Elastic-Net regression, see sklearn.linear_model.ElasticNet, or sklearn.linear_model.enet_path, and pass the parameters fit_intercept=False, positive=True
User Guide See the Nonnegative Least Squares Regression section for further details.
Classes¶
regressors.NonnegativeLinear |
Nonnegative least squares linear model. |
regressors.NonnegativeRidge ([alpha, solver]) |
Nonnegative least squares with L2 regularization. |
Functions¶
regressors.nonnegative_regression (X, y[, …]) |
Solve the nonnegative least squares estimate regression problem. |
regressors.nonnegative_ridge_regression (X, …) |
Solve the nonnegative least squares estimate ridge regression problem. |
mcmodels.regressors.nonparametric
: Nonparametric Regression¶
The mcmodels.regressors.nonparametric
module implements Nonparametric
regression models and the polynomial family of kernels. This module includes
NadarayaWatson regression and the general Polynomial kernel.
User Guide See the Nonnparametric Regression section for further details.
regressors.NadarayaWatson ([kernel, degree, …]) |
NadarayaWatson Estimator. |
regressors.NadarayaWatsonCV (param_grid[, …]) |
NadarayaWatson Estimator with built in Leave-one-out cross validation. |
Kernels:¶
regressors.nonparametric.kernels.Polynomial ([…]) |
Polynomial kernel. |
regressors.nonparametric.kernels.Uniform ([…]) |
Uniform kernel. |
regressors.nonparametric.kernels.Epanechnikov ([…]) |
Epanechnikov kernel. |
regressors.nonparametric.kernels.Biweight ([…]) |
Biweight kernel. |
regressors.nonparametric.kernels.Triweight ([…]) |
Triweight kernel. |
mcmodels.models
: Published Mesoscale Connectivity Models¶
The mcmodels.models
module implements models that have been developed
here at the Allen Institute for modeling mesoscale connectivity in the mouse.
The module contains a HomogeneousModel similar to Oh et al. 2014 as well as the
recent VoxelModel from Knox et al. 2018.
mcmodels.models.homogeneous
: Homogeneous Regional Model¶
The mcmodels.models.homogeneous
module implements a Homogeneous
Model for predicting regional connectivity similar to Oh et al. 2014.
Additionally, the module implements greedy forward/backward subset selection
algorithms to improve the conditioning of given input arrays.
User Guide See the HomogeneousModel similar to [Oh2014] section for further details.
Classes¶
models.HomogeneousModel ([kappa]) |
Homogeneous model similar to Oh et al. |
Functions¶
models.homogeneous.svd_subset_selection (X, n) |
svd subset selection to return n cols that are less linearly dependent. |
models.homogeneous.forward_subset_selection_conditioning (X) |
Conditioning through subselecting columns of X. |
models.homogeneous.backward_subset_selection_conditioning (X) |
Conditioning through subselecting columns of X. |
mcmodels.models.voxel
: Voxel-scale Model¶
The mcmodels.models.voxel
module implements the voxel scale model
or predicting voxel-scale connectivity from Knox et al. 2018.
The module also contains a VoxelConnectivityArray class to work with the model
in memory (implicitly computing its weights) and a RegionalizedModel class to
integrate the model weights to the regional level.
User Guide See the New Voxel-scale Connectivity Model [Knox2018] section for further details.
models.VoxelModel (source_voxels[, kernel, …]) |
Voxel-scale interpolation model for mesoscale connectivity. |
models.voxel.RegionalizedModel (weights, …) |
Regionalization/Parcelation of VoxelModel. |
models.voxel.VoxelConnectivityArray (weights, …) |
Class for implicit construction of the voxel model. |
mcmodels.utils
: Utilities¶
Module containing utility functions
utils.get_experiment_ids (mcc, structure_ids) |
Returns all experiment ids with injection in structure_ids. |
utils.lex_ordered_unique (ar, lex_order[, …]) |
np.unique in a given order. |
utils.nonzero_unique (ar, \*\*unique_kwargs) |
np.unique returning only nonzero unique elements. |
utils.ordered_unique (ar, \*\*unique_kwargs) |
np.unique in the order in which the unique values occur. |
utils.padded_diagonal_fill (arrays) |
Returns array filled with uneven arrays padding with zeros. |
utils.unionize (volume, key[, return_regions]) |
Unionize voxel data to regional data. |