Build Model Map

class Orange.modelmaps.BuildModelMap(fname, folds=10, model_limit=500, seed=42)

build_model(attributes, learner)

Build a classification meta-model.

Parameters:

• attributes (list of strings) – subset of attributes
• learner (Orange.classification.Learner) – classification learner to wrap

build_model_data(models)

Return an Orange.data.Table of model meta-data.

build_model_matrix(models, dist=<function distance_manhattan at 0x3735b90>)

Build a distance matrix of models given the distance measure.

build_projection_model(attributes, viz_method)

Build a projection meta-model.

Parameters:

• attributes (list of strings) – attributes for projection
• viz_method (enum) – visualization method

build_rf_models(trees=50, max_depth=4, min_instances=5)

Build Random forest and return tree models.

Parameters:

• trees (int) – number of trees in the forest
• max_depth (int) – maximal tree depth
• min_instances (int) – nodes with less than min_instances instances are not split further

select_representatives(models, dist=<function distance_euclidean at 0x3735c08>)

Contruct a network, detect communities and return representatives.

Parameters:

• models (list of modelmaps.Model) – select representatives from this models
• dist (func) – distance function

Help Functions

Orange.modelmaps.modelmap.load(file_name)

Load model map.

Read compressed tuple containing model similarity matrix and data table.

Orange.modelmaps.modelmap.save(file_name, smx, model_data, original_data)

Save model map.

Model similarity matrix and data table tuple is pickled and compressed as a bz2 archive.

Orange.modelmaps.modelmap.get_models_table()

Return an empty data table for model meta data.

Orange.modelmaps.modelmap.get_feature_subsets(domain, nsubsets=None, min_features=None, max_features=None, seed=None)

Return random attribute subsets.

Parameters:

• domain (Orange.data.Domain) – data set domain to extract features
• nsubsets (int) – number of attribute subsets

Orange.modelmaps.modelmap.model_network(smx, table=None, knn=1)

Build network from model distance matrix.

Parameters:

• smx (Orange.misc.SymMatrix) – model distance matrix
• table (Orange.data.Table) – model meta data
• knn (int) – connect each model with knn neighbours