neuralnetworknode properties

Last updated: Jan 18, 2024

Neural Net node icon The Neural Net node uses a simplified model of the way the human brain processes information. It works by simulating a large number of interconnected simple processing units that resemble abstract versions of neurons. Neural networks are powerful general function estimators and require minimal statistical or mathematical knowledge to train or apply.

Example

node = stream.create("neuralnetwork", "My node")
# Build Options tab - Objectives panel
node.setPropertyValue("objective", "Standard")
# Build Options tab - Ensembles panel
node.setPropertyValue("combining_rule_categorical", "HighestMeanProbability")

Table 1. neuralnetworknode properties
`neuralnetworknode` Properties	Values	Property description
`targets`	[field1 ... fieldN]	Specifies target fields.
`inputs`	[field1 ... fieldN]	Predictor fields used by the model.
`splits`	[field1 ... fieldN	Specifies the field or fields to use for split modeling.
`use_partition`	flag	If a partition field is defined, this option ensures that only data from the training partition is used to build the model.
`continue`	flag	Continue training existing model.
`objective`	`Standard` `Bagging` `Boosting` `psm`	`psm` is used for very large datasets, and requires a server connection.
`method`	`MultilayerPerceptron` `RadialBasisFunction`
`use_custom_layers`	flag
`first_layer_units`	number
`second_layer_units`	number
`use_max_time`	flag
`max_time`	number
`use_max_cycles`	flag
`max_cycles`	number
`use_min_accuracy`	flag
`min_accuracy`	number
`combining_rule_categorical`	`Voting` `HighestProbability` `HighestMeanProbability`
`combining_rule_continuous`	`MeanMedian`
`component_models_n`	number
`overfit_prevention_pct`	number
`use_random_seed`	flag
`random_seed`	number
`missing_values`	`listwiseDeletion` `missingValueImputation`
`use_model_name`	boolean
`model_name`	string
`confidence`	`onProbability` `onIncrease`
`score_category_probabilities`	flag
`max_categories`	number
`score_propensity`	flag
`use_custom_name`	flag
`custom_name`	string
`tooltip`	string
`keywords`	string
`annotation`	string
`calculate_variable_importance`	boolean	For models that produce an appropriate measure of importance, you can display a chart that indicates the relative importance of each predictor in estimating the model. Typically, you'll want to focus your modeling efforts on the predictors that matter most, and consider dropping or ignoring those that matter least.