*ML_LEARN_RFC

Description: Random Forest Classifier

Syntax: ml_learn_rfc(dataset,independents,clabel,cv,train_ratio,normalize,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
5	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
6	Normalize	select	none	Normalizes each numerical column.
7	Auto Clean	select	no	Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_DTREE

Description: Decision Tree Classifier

Syntax: ml_learn_dtree(dataset,independents,clabel,criterion,cv,train_ratio,normalize,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Criterion	select	gini	The method used to measure the quality of the splits in a tree.
5	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
6	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
7	Normalize	select	none	Normalizes each numerical column.
8	Auto Clean	select	no	Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_ELASTICNET

Description: ElasticNet Regression

Syntax: ml_learn_elasticnet(dataset,independents,clabel,normalize,autoclean,grid_response,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
11	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_BAYESIANRIDGE

Description: BayesianRidge Regression

Syntax: ml_learn_bayesianridge(dataset,independents,clabel,normalize,autoclean,grid_response,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
11	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_BAYESIANRIDGECV

Description: BayesianRidge Regression with CV

Syntax: ml_learn_bayesianridgecv(dataset,independents,clabel,normalize,autoclean,grid_response,cv,train_ratio,dependent_targets,num_dig)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_DECISION_TREE_TRAIN

Description: Train or classify using Decision Tree

Syntax: ml_decision_tree_train(database_id,train_column,drop_columns)

Inputs

ID	Name	Type	Default	Remarks
1	Database	text
2	Column Name	file
3	Drop Columns	scalar

Outputs

ID	Name	Type	Remarks
1		text

*ML_LEARN_LOGISTIC

Description: Logistic Regression

Syntax: ml_learn_logistic(dataset,independents,clabel,normalize,autoclean,grid_response,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
11	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_SGDR

Description: SGD Regression

Syntax: ml_learn_sgdr(dataset,independents,clabel,degree,normalize,grid_response,autoclean,cv,train_ratio,optimize,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Degree	select	none	Order of fit.
5	Normalize	select	none	Normalizes each numerical column.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Auto Clean	select	no	Clean a dataset using preset rules.
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Optimize	select		Obtain the input points and resulting Optimum.
11	Use Target Dependency	select	no	Useful for time-history predictions
12	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
13	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_LASSO

Description: Lasso Regression

Syntax: ml_learn_lasso(dataset,independents,clabel,degree,normalize,grid_response,autoclean,run_grid_search,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Degree	select	none	Order of fit.
5	Normalize	select	none	Normalizes each numerical column.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Auto Clean	select	no	Clean a dataset using preset rules.
8	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
9	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
10	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
11	Use Target Dependency	select	no	Useful for time-history predictions
12	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
13	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_RIDGE

Description: Ridge Regression

Syntax: ml_learn_ridge(dataset,independents,clabel,normalize,grid_response,autoclean,degree,run_grid_search,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Degree	select	none	Order of fit.
8	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
9	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
10	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
11	Use Target Dependency	select	no	Useful for time-history predictions
12	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
13	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_DTREEREG

Description: Decision Tree Regression

Syntax: ml_learn_dtreereg(dataset,independents,clabel,normalize,grid_response,autoclean,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
11	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_GPR

Description: Gaussian Process Regression

Syntax: ml_learn_gpr(dataset,independents,clabel,normalize,grid_response,autoclean,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
11	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_GBOOSTR

Description: Gradient Boost Regressor

Syntax: ml_learn_gboostr(dataset,independents,clabel,normalize,grid_response,autoclean,run_grid_search,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Use Target Dependency	select	no	Useful for time-history predictions
11	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
12	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_RFR

Description: Random Forest Regression

Syntax: ml_learn_rfr(dataset,independents,clabel,normalize,grid_response,autoclean,run_grid_search,cv,train_ratio,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Use Target Dependency	select	no	Useful for time-history predictions
11	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
12	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_SVR

Description: Support Vector Regression

Syntax: ml_learn_svr(dataset,independents,clabel,normalize,grid_response,autoclean,cv,train_ratio,dependent_targets,num_dig)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
8	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
9	Use Target Dependency	select	no	Useful for time-history predictions
10	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_MLP

Description: MLP Regression

Syntax: ml_learn_mlp(dataset,independents,clabel,normalize,grid_response,autoclean,run_grid_search,cv,train_ratio,dependent_targets,mlp_hidden_layer_sizes,mlp_max_iter,mlp_alpha,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
6	Auto Clean	select	no	Clean a dataset using preset rules.
7	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Use Target Dependency	select	no	Useful for time-history predictions
11	MLP Hidden Layer Sizes (csv)	text		MLP Hidden layers separarted by commas. Each delimited values represents a layer between the Input and the Output layers
12	MLP Maximum Number of Iterations	text		MLP Hidden layers separarted by commas. Each delimited values represents a layer between the Input and the Output layers
13	MLP Strength of the L2 regularization term	text		Strength of the L2 Regularization term
14	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
15	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_MLP_CLASSIFY

Description: MLP-Classifier

Syntax: ml_learn_mlp_classify(dataset,independents,clabel,train_ratio,normalize,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
5	Normalize	select	none	Normalizes each numerical column.
6	Auto Clean	select	no	Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_GNB

Description: Gaussian Naive Bayes

Syntax: ml_learn_gnb(dataset,independents,clabel,normalize,train_ratio,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
6	Auto Clean	select	no	Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_PREDICT_INTERACTIVE

Description: ML Predict Info

Syntax: ml_predict_interactive(mfile,inputs,targets,schema,dataset,reference_dataset,raw_column_name)

Inputs

ID	Name	Type	Default	Remarks
1	Model File	text		Trained model file
2	Inputs	scalar		This can come from a ML_PREDICT_INFO worker that were used while learning
3	Targets	scalar		This can come from a ML_PREDICT_INFO worker that were used while learning
4	Schema	dataset		This can come from a ML_PREDICT_INFO worker that were used while learning <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
5	Inputs For Prediction	predict_interactive		If the inputs, targets and schema are specified, a slider will be presented for selection
6	Reference Dataset	dataset		This dataset contains the points that will be original dataset with curves <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
7	Raw Column Name	scalar		If the prediction is for curve points, you can choose the Curve Column name during training

Outputs

ID	Name	Type	Remarks
1	Predictions	dataset

*ML_PREDICT_INFO

Description: ML Interactive Prediction

Syntax: ml_predict_info(mfile)

Inputs

ID	Name	Type	Default	Remarks
1	Saved Model File	text		Path of the model file from a previous training worker

Outputs

ID	Name	Type	Remarks
1	Model File	text
2	Lucy JSON	json
3	Independents	scalar
4	Targets	scalar
5	Schema	dataset
6	Config Parameters	keyvalue

*ML_PREDICT

Description: ML Predict

Syntax: ml_predict(dataset,mfile,reference_dataset,raw_curve_column)

Inputs

ID	Name	Type	Default	Remarks
1	Prediction Dataset	dataset		Input Dataset for Prediction <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
2	Saved Model File	text
3	Learn Dataset	dataset		Useful for curve predictions <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
4	Raw Curve Column Name	text		Useful for curve predictions so we can use the x-values of this curve to build the Predicted curve

Outputs

ID	Name	Type	Remarks
1	Predictions	dataset

*ML_LEARN_SVC

Description: Support Vector Classifier

Syntax: ml_learn_svc(dataset,independents,clabel,cv,train_ratio,normalize,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
5	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
6	Normalize	select	none	Normalizes each numerical column.
7	Auto Clean	select	no	Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_AUTO

Description: Auto Regression

Syntax: ml_learn_auto(dataset,independents,clabel,regression_types,normalize,grid_response,autoclean,cv,train_ratio,run_grid_search,dependent_targets,num_dig,skip_trainingdata_in_model,gs_scoring_type,cv_scoring_type,ref_dataset,ref_ratio)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Regression Type	select	Array	Available regression methods
5	Normalize	select	none	Normalizes each numerical column.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Auto Clean	select	no	Clean a dataset using preset rules.
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
11	Use Target Dependency	select	no	Useful for time-history predictions
12	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
13	Skip including training data in model	select	no	While saving the model, we skip including the training data
14	Grid Search Scoring Type	select	r2	How to measure error
15	Cross-Validation Scoring Type	select	r2	How to measure error
16	Verification Dataset	dataset		The trained model will be tested using this dataset
17	Hold Back Ratio	scalar	0	When this ratio is non-zero, the model will be trained on the dataset and tested on the verification dataset

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_IMAGE_CLASSIFICATION_LEARN

Description: Connects to a d3VIEW Image Classification Server to learn based on annotations generated from Image Annotator

Syntax: ml_image_classification_learn(base_url,dataset,config)

Inputs

ID	Name	Type	Default	Remarks
1	Base URL	text		Url of the server running d3VIEW Image classification
2	LDWA	dataset		This would be a set of manually annotated images generated by the image annotation utility worker in Workflows <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
3	Configuration Options	keyvalue		Configuration Options <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#keyvalueinput’> <i class=’fa fa-external-link’> </i> View more </a>

Outputs

ID	Name	Type	Remarks
1	UID	text
2	Model Path	textarea
3	Base Url	text
4	Status Url	text
5	Logs	dataset

*ML_SIMLYTIKS_TERMINAL

Description: Learns and predicts the commands for Simlytiks based on data

Syntax: ml_simlytiks_terminal(server_url,dataset,schema,visualizations,type,uid,query)

Inputs

ID	Name	Type	Default	Remarks
1	Server URL	text		Url of the server running d3VIEW Terminal Server
2	Dataset	dataset		Input Dataset <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
3	Schema	dataset		Schema of the dataset that contains the column information <a class=’btn btn-xs btn-default’ target=’_blank’ href=’https://www.d3view.com/docs/master/workflows/Glossary.html#datasetinput’> <i class=’fa fa-external-link’> </i> View more </a>
4	Visualizations	json		JSON file that contains the visualization information
5	Learn Or Predict	select		Learn or Predict for the terminal
6	UID	text		Unique Identifier to use while predicting
7	Predict Query	textarea		Text to predict Simlytiks command

Outputs

ID	Name	Type	Remarks
1	UID	text
2	Prediction	json

*ML_LEARN_MEANSHIFT

Description: Mean Shift Clustering

Syntax: ml_learn_meanshift(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_AFFINITYPROPAGATION

Description: Affinity Propagation

Syntax: ml_learn_affinitypropagation(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_EXPLORE_RUN_PCA

Description: RunPCA

Syntax: ml_explore_run_pca(dataset,independents,normalize)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Normalize	select	none	Normalizes each numerical column.

Outputs

ID	Name	Type	Remarks
1	Lucy-ML JSON output	json

*ML_EXPLORE_GET_SCHEMA

Description: Get Schema

Syntax: ml_explore_get_schema(dataset,independents,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Auto Clean	select		Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy-ML JSON output	json

*ML_LEARN_AGGLOMERATIVECLUSTERING

Description: Agglomerative Clustering

Syntax: ml_learn_agglomerativeclustering(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_SPECTRALCLUSTERING

Description: Spectral clustering

Syntax: ml_learn_spectralclustering(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_LINEAR

Description: Linear Regression

Syntax: ml_learn_linear(dataset,independents,clabel,degree,normalize,grid_response,autoclean,cv,train_ratio,run_grid_search,dependent_targets,num_dig,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Degree	select	none	Order of fit.
5	Normalize	select	none	Normalizes each numerical column.
6	Generate Regression Line/Surface	select	yes	Used to visualize a regression line or surface.
7	Auto Clean	select	no	Clean a dataset using preset rules.
8	Selection of Validation	select	leave_one_out	Ability to valide the model using different methods. The default is test_train_split
9	Train ratio	text	0.9	Percentage of data to be used for training. Remaining will be used for testing
10	Run Grid Search	select	no	Runs grid search to find the best hyperparameters
11	Use Target Dependency	select	no	Useful for time-history predictions
12	Curve Digitize Points	select	none	Number of digitized points to use when the Input or the Target column is a curve
13	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_BIRCHCLUSTERING

Description: Birch clustering

Syntax: ml_learn_birchclustering(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_LEARN_DBSCANCLUSTERING

Description: DBSCAN clustering

Syntax: ml_learn_dbscanclustering(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_EXPLORE_FEATURE_IMPORTANCE

Description: Feature Importance

Syntax: ml_explore_feature_importance(dataset,independents,clabel,normalize,autoclean)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Target Column	select		The name of the target column.
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select		Clean a dataset using preset rules.

Outputs

ID	Name	Type	Remarks
1	Lucy-ML JSON output	json

*ML_LEARN_KMEANS

Description: K-Means

Syntax: ml_learn_kmeans(dataset,independents,nc,normalize,autoclean,skip_trainingdata_in_model)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset
2	Input Features	select		A subset of columns that will be used to train the model.
3	Number of Clusters	text	3	Total number of clusters to group the data into. If set to auto, an optimum cluster size will be found based on distance optimization
4	Normalize	select	none	Normalizes each numerical column.
5	Auto Clean	select	no	Clean a dataset using preset rules.
6	Skip including training data in model	select	no	While saving the model, we skip including the training data

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Responses from Learn	dataset
3	Visualizations from Learn	dataset
4	Score	number
5	Model Path	textarea

*ML_CLEAN_AUTOCLEAN

Description: Auto Clean

Syntax: ml_clean_autoclean(dataset)

Inputs

ID	Name	Type	Default	Remarks
1	Dataset	dataset

Outputs

ID	Name	Type	Remarks
1	Lucy ML JSON Output	json
2	Cleaned Dataset	dataset