On this Page

Overview

This is a Transform type Snap that builds model for a classification dataset. In the Snap's settings, you can select the target field in the dataset, algorithm, and configure parameters for the selected algorithm.

If you want to build the model on regression dataset, use Trainer -- Regression Snap instead.

Input and Output

Expected input: The classification dataset.

Expected output: A serialization of the model, and metadata which are not human-readable. Additionally, the output includes a human-readable representation of the model if the Readable checkbox is selected.

Expected upstream Snap: Any Snap that generates a classification dataset document. For example, CSV Generator, JSON Generator, or a combination of File Reader and JSON Parser.

Expected downstream Snap: Snaps that require a model input. For example, the Predictor (Classification) Snap. Alternatively, any Snap that stores the model to be used in another pipeline. For example, a combination of JSON Formatter and File Writer.

Prerequisites

The data from upstream Snap must be in tabular format (no nested structure).
This Snap automatically derives the schema (field names and types) from the first document. Therefore, the first document must not have any missing values.

Configuring Accounts

Accounts are not used with this Snap.

Configuring Views

Input	This Snap has exactly one document input view.
Output	This Snap has exactly one document output view.
Error	This Snap has at most one document error view.

Troubleshooting

None.

Limitations and Known Issues

None.

Modes

Ultra Pipelines: Does not work in Ultra Pipelines.

Snap Settings

Label	Required. The name for the Snap. Modify this to be more specific, especially if there are more than one of the same Snap in the pipeline.
Label field	Required. The label or output field in the dataset. This must be a categorical type represented as text (string data type). This is the field that the model will be trained to predict. Default value: None Example: $class
Algorithm	Required. The classification algorithm to be used to build the model. There are eight classification algorithms currently available: Decision Tree K-Nearest Neighbors Logistic Regression Naive Bayes Support Vector Machines Decision Stump Random Forests Multilayer Perceptron The implementations are from WEKA, an open source machine learning library in Java. Default value: Decision Tree
Options	The parameters to be applied on the selected classification algorithm. Each algorithm has a different set of parameters to be configured in this property. If this property is left blank, the default values are applied for all the parameters. If specifying multiple parameters, separate them with a comma ",". See Options for Algorithms section below for details. Default value: None Examples: batch_size = 120 batch_size = 120, collapse_tree = true
Readable	Select this to output the model in human-readable format. When selected, a $readable field is added to the output, this displays the model in a readable format. Default value: Not selected
Snap Execution	Select one of the following three modes in which the Snap executes: Validate & Execute: Performs limited execution of the Snap, and generates a data preview during Pipeline validation. Subsequently, performs full execution of the Snap (unlimited records) during Pipeline runtime. Execute only: Performs full execution of the Snap during Pipeline execution without generating preview data. Disabled: Disables the Snap and all Snaps that are downstream from it. Default Value: Execute only Example: Validate & Execute

Best Practices

Using Type Converter Snap Upstream

In some cases, the numerical fields may be represented as text. You can use the Type Converter Snap to convert data into appropriate types before feeding into the Trainer (Classification) Snap.

Algorithm Selection

In order to choose the best possible algorithm for your dataset, use the Cross Validator (Classification) Snap to perform k-fold cross validation on the dataset. The algorithm that produces the best accuracy is likely to be the one most suitable for your dataset. Apply the same algorithm for your dataset in the Trainer (Classification) Snap to build the model.

Options for Algorithms

Options for algorithms

This section lists all applicable parameters for each algorithm along with their default values.

Decision Tree

Option Name	Option Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
binary_splits	boolean	false	Enable binary splits.
collapse_tree	boolean	true	Set to allow collapsing the tree.
confidence_factor	float	0.25	Set the confidence threshold for pruning.
do_not_check_capabilities	boolean	false	Enable capability-check.
do_not_mask_split_point_actual_value	boolean	false	Set whether the split point actual value is to be masked.
min_num_obj	int	2	Set the minimum number of instances per leaf.
num_decimal_places	int	2	Set the number of decimal places.
num_folds	int	3	Set the number of folds for reduced error pruning. One fold is used as pruning set.
reduced_error_pruning	boolean	false	Enable reduced error pruning. This is false if the unpruned parameter is set to true.
save_instance_data	boolean	false	Set whether the instance data is to be saved. If set to true, it does not clean up after the tree has been built.
seed	int	1	Set the seed for random data shuffling.
subtree_raising	boolean	true	Enable sub-tree raising.
unpruned	boolean	false	Enable using unpruned tree. This is false if the reduced_error_pruningparameter is set to true.
use_laplace	boolean	false	Allow Laplace smoothing for predicted probabilities.
use_md_correction	boolean	true	Allow MDL correction for info gain on numeric attributes.

K-Nearest Neighbors

Option Name	Option Data Type	Default Value	Description
knn	int	1	Set the number of neighbors the learner will use.
batch_size	int	100	Set the preferred batch size for batch prediction.
cross_validate	boolean	false	Set whether the hold-one-out cross-validation will be used to select the best k-value.
distance_weighting	enum [weight_none, weight_inverse, weight_similarity]	weight_none	Set the distance weighting method used.
do_not_check_capabilities	boolean	false	Enable capability-check.
mean_squared	boolean	false	Set whether the mean squared error is used rather than mean absolute error when doing cross-validation.
nearest_neighbours_search_algorithm	enum [LinearNNSearch, BallTree, CoverTree, KDTress, FilteredNeighbourSearch]	LinearNNSearch	Set the nearestNeighbourSearch algorithm to be used for finding nearest neighbor(s).
num_decimal_places	int	2	Set the number of decimal places.
window_size	int	0 (no window size)	Set the maximum number of instances allowed in the training pool.

Logistic Regression

Option Name	Option Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
do_not_check_capabilities	boolean	false	Enable capability-check.
matx_its	int	-1	Set the maximum number of iterations.
num_decimal_places	int	4	Set the number of decimal places.
ridge	double	1.00E-08	Sets the ridge in the log-likelihood.
use_conjugate_gradient_descent	boolean	false	Sets whether conjugate gradient descent is used.

Naive Bayes

Option Name	Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
do_not_check_capabilities	boolean	false	Enable capability-check.
num_decimal_places	int	2	Set the number of decimal places.
use_kernel_estimator	boolean	false	Use kernel density estimator rather than normal distribution for numeric attributes. This is false if use_supervised_discretization is set to true.
use_supervised_discretization	boolean	false	Use supervised discretization to process numeric attributes. This is false if use_kernel_estimator is set to true.

Support Vector Machines

Option Name	Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
c	double	1	Set the complexity constant C.
build_calibration_models	boolean	false	Set whether to fit calibration models to SVM outputs.
checks_turned_off	boolean	false	Disable all checks.
do_not_check_capabilities	boolean	false	Enable capability-check.
epsilon	double	1.00E-12	Set the epsilon for round-off error.
filter_type	enum [filter_none, filter_normalize, filter_standardize]	filter_normalize	Set how the training data will be transformed.
kernel	enum [PolyKernel, NormalizedPolyKernel, PrecomputedKernelMatrixKernel, Puk, RBFKernel, StringKernel]	PolyKernel	Specify the Kernel to use.
num_decimal_places	int	2	Set the number of decimal places.
num_folds	int	-1	Set the number of folds for the internal cross-validation.
random_seed	int	1	Set the random number seed.
tolerance_parameter	double	1.00E-03	Set the tolerance parameter.

Decision Stump

Option Name	Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
do_not_check_capabilities	boolean	false	Enable capability-check.
num_decimal_places	int	2	Set the number of decimal places.

Random Forests

Option Name	Data Type	Default Value	Description
bag_size_percent	int	100	Specify the size of each bag, as a percentage of the training set size.
batch_size	int	100	Set the preferred batch size for batch prediction.
break_ties_randomly	boolean	false	Set whether to break ties randomly when several attributes look equally good.
calc_out_of_bag	boolean	false	Set whether to calculate the out-of-bag error.
compute_attribute_importance	boolean	false	Set whether compute and output attribute importance (mean impurity decrease method)
do_not_check_capabilities	boolean	false	Enable capability-check.
max_depth	int	0	Set the maximum depth of the tree, 0 for unlimited.
num_decimal_places	int	2	Set the number of decimal places.
num_execution_slots	int	1	Set the number of execution slots.
num_features	int	0	Set the number of attributes to randomly investigate.
num_iterations	int	100	Set the number of iterations.
seed	int	1	Set the seed for random number generator.
store_out_of_bag_predictions	boolean	false	Set whether to store out-of-bag predictions in internal evaluation object.

Multilayer Perceptron

Option Name	Data Type	Default Value	Description
batch_size	int	100	Set the preferred batch size for batch prediction.
do_not_check_capabilities	boolean	false	Enable capability-check.
learning_rate	double	0.3	Set the learning rate for the backward propagation. The value should be between 0 - 1.
momentum	double	0.2	Set the momentum rate for the backward propagation algorithm.The value should be between 0 - 1.
num_decimal_places	int	2	Set the number of decimal places.
seed	int	0	Set the value used to seed the random number generator.
validation_set_size	int	0	Set the percentage size of validation set to use to terminate training. The value should be between 0 - 1.
validation_threshold	int	20	Set the number of consecutive increases of error allowed for validation testing before training terminates.

Example

Weight Balance Classification – Model Training

This pipeline demonstrates training a model to predict whether a weighing scale is balanced. The classification algorithm is selected based on the algorithm evaluation in the Cross Validator (Classification) Snap's example. The input dataset depicts the weight on each side of the scale and the side's distance from the floor.

Download this pipeline.

Understanding the pipeline

Input

The input document is generated by the CSV Generator Snap and is composed of four fields, one classification field and three numeric fields:

Balance Class: The classification field to denote status of the weighing scale. B for Balanced, L for Left-inclined, and R for Right-inclined.
Left Weight
Left Distance
Right Weight
Right Distance

This dataset has been sourced from UCI Dataset Archive.

Objective

Use Trainer (Classification) Snap to train the model for the dataset.

Data Preparation

This input document is passed through the Type Converter Snap that is configured to automatically detect and convert the data types. In any ML pipeline, you must first analyze the input document using the Profile Snap and the Type Inspector Snap to ensure that there are no null values or that the data types are accurate. This step is skipped in this example for simplicity's sake.

Model Training

Since the training algorithm was evaluated in the Cross Validator (Classification) Snap, the Trainer (Classification) Snap is configured with the same settings:

The output from this Snap is as shown below:

The model in the Trainer (Classification) Snap's output is written into a file using the File Writer Snap which is configured as shown below:

This model is used to predict the Balance Class for an unlabeled dataset. See Weight Balance Classification – Testing for details.