One AI Machine Learning Algorithms and Settings

One AI Machine Learning is an AutoML pipeline, meaning it’s an automated process for building machine learning models.  One of the most powerful things this pipeline does is produce the best performing model with the least amount of user intervention possible.  To accomplish this, a number of models are generated using different algorithms and settings each time a machine learning augmentation is run.  The best performing model is then leveraged to make the predictions.  This default behavior can be overridden by selecting a specific algorithm or algorithms in the configuration settings prior to running.  

What follows is a list of all of the algorithms and settings attempted by default as well as the additional algorithms available.  Where multiple values are listed for specific settings, multiple models are compared unless that setting is overridden.

Classification Algorithms

Classifiers Attempted by Default

LGBMClassifier https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMClassifier.html 

boosting_type: 'gbdt', 'goss'

n_estimators: 100

reg_lambda: 0.0

num_leaves: 31

learning_rate: 0.1

reg_alpha: 0.0

subsample: 1.0

subsample_freq: 0

colsample_bytree: 1.0

verbosity: -1

min_child_samples: 20

objective: None

class_weight: None

importance_type: 'split'

n_jobs: -1

max_depth: -1

subsample_for_bin: 200000

min_child_weight: 0.001

min_split_gain: 0.0

    

AdaBoostClassifier https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostClassifier.html 

n_estimators: 5

learning_rate: 1.0

algorithm: "SAMME.R"

base_estimator: None

    

RandomForestClassifier https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html 

n_jobs: -1

criterion: 'gini'

max_features: 'auto', 'sqrt', 'log2'

max_depth: None, 4, 8, 10, 12

min_samples_split: 2

min_samples_leaf: 1

min_weight_fraction_leaf: 0

max_leaf_nodes: None

min_impurity_decrease: 0

bootstrap: True

oob_score: False

verbose: 0

warm_start: False

class_weight: 'balanced'

n_estimators: 5

ccp_alpha: 0.0

max_samples: None

 

LogisticRegression https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html 

penalty: 'l1', 'l2'

dual: False

tol: 1e-4

C: 0.5, 1.0, 2.0

fit_intercept: True, False

intercept_scaling: 1.0

class_weight: None, 'balanced'

solver: 'liblinear'

max_iter: 100

multi_class: 'ovr'

verbose: 0

warm_start: False

l1_ratio: None

n_jobs: None

    

GaussianNB http://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.GaussianNB.html 

var_smoothing: 1e-9

priors: None

Optional Classifiers    

DecisionTreeClassifier https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html 

criterion: "gini"

splitter: 'best'

max_depth: 4, 5

min_samples_split: 2, 3, 4, 5

min_samples_leaf: 1, 2, 4

min_weight_fraction_leaf: 0

max_features: None

max_leaf_nodes: None

min_impurity_decrease: 0

class_weight: None, 'balanced'

ccp_alpha: 0.0

    

KNeighborsClassifier https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html 

weights: 'uniform'

algorithm: 'auto'

leaf_size: 30

p: 2

metric: 'minkowski'

metric_params: None

n_jobs: -1

n_neighbors: 5

 

SVC https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html 

kernel: 'linear', 'poly', 'rbf', 'sigmoid'

gamma: 'auto'

shrinking: True, False

max_iter: -1

cache_size: 200

coef0: 0.0

decision_function_shape: 'ovr'

degree: 3

verbose: False

tol: 0.001

C: 1.0

class_weight: None

probability: False

break_ties: False

    

GaussianProcessClassifier https://scikit-learn.org/stable/modules/generated/sklearn.gaussian_process.GaussianProcessClassifier.html 

warm_start: False

multi_class: 'one_vs_rest'

n_jobs: None

optimizer: 'fmin_l_bfgs_b'

kernel: None

n_restarts_optimizer: 0

max_iter_predict: 100

copy_X_train: True

Regression Algorithms

Regressors Attempted by Default

LinearRegression https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html 

fit_intercept: True, False

normalize: False

copy_X: True

n_jobs: None

positive: False

 

RandomForestRegressor https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html 

n_estimators: 10

criterion: 'mse'

max_depth: None

min_samples_split: 2

min_samples_leaf: 1

min_weight_fraction_leaf: 0.0

max_features: 'auto'

max_leaf_nodes: None

bootstrap: True

min_impurity_decrease: 0

oob_score: False

n_jobs: None

verbose: 0

warm_start: False

ccp_alpha: 0.0

max_samples: None

 

Lasso https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Lasso.html 

alpha: 1.0

fit_intercept: False

normalize: False

precompute: False

copy_X: True

max_iter: 1000

tol: 0.0001

warm_start: False

positive: False

selection: 'cyclic'

    

ElasticNet https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.ElasticNet.html 

alpha: 1.0

l1_ratio: 0.5

fit_intercept: False

normalize: False

precompute: False

max_iter: 1000

copy_X: True

tol: 0.0001

warm_start: False

positive: False

selection: 'cyclic'

    

LGBMRegressor https://lightgbm.readthedocs.io/en/latest/pythonapi/lightgbm.LGBMRegressor.html 

boosting_type: 'gbdt', 'goss'

n_estimators: 100

reg_lambda: 0.0

num_leaves: 31

learning_rate: 0.1

reg_alpha: 0.0

subsample: 1.0

subsample_freq: 0

colsample_bytree: 1.0

silent: True

min_child_samples: 20

objective: None

importance_type: 'split'

n_jobs: -1

max_depth: -1

subsample_for_bin: 200000

min_child_weight: 0.001

min_split_gain: 0.0

Optional Regressors

SVR https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVR.html 

kernel: 'rbf'

degree: 3

gamma: 'auto'

coef0: 0.0

tol: 0.001

C: 1.0

epsilon: 0.1

shrinking: True

cache_size: 200

verbose: False

max_iter: -1

    

Ridge https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.ridge_regression.html 

alpha: 1.0

fit_intercept: True

normalize: False

copy_X: True

max_iter: None

tol: 0.001

solver: 'auto'

positive: False

    

HuberRegressor https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.HuberRegressor.html 

epsilon: 1.35

max_iter: 100

alpha: 0.0001

warm_start: False

fit_intercept: True

tol: 1e-05

    

GaussianProcessRegressor https://scikit-learn.org/stable/modules/generated/sklearn.gaussian_process.GaussianProcessRegressor.html 

kernel: None

alpha: 1e-10

optimizer: 'fmin_l_bfgs_b'

n_restarts_optimizer: 0

normalize_y: False

copy_X_train: True

 

SGDRegressor https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDRegressor.html 

loss: 'squared_loss'

penalty: 'l2'

alpha: 0.0001

l1_ratio: 0.15

fit_intercept: True

max_iter: 1000

tol: 0.001

shuffle: True

verbose: 0

epsilon: 0.1

learning_rate: 'invscaling'

eta0: 0.01

power_t: 0.25

early_stopping: False

validation_fraction: 0.1

n_iter_no_change: 5

warm_start: False

average: False    

 

DecisionTreeRegressor https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeRegressor.html 

criterion: "squared_error"

splitter: 'best'

max_depth: 4, 5

min_samples_split: 2, 3, 4, 5

min_samples_leaf: 1, 2, 4

min_weight_fraction_leaf: 0

max_features: None

max_leaf_nodes: None

min_impurity_decrease: 0

ccp_alpha: 0.0

    

AdaBoostRegressor https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostRegressor.html 

n_estimators: 50

learning_rate: 1.0

loss: 'linear','square'

Was this article helpful?

0 out of 0 found this helpful

Comments

0 comments

Please sign in to leave a comment.