Frequently Asked Questions

550 questions found.

Page 11 / 11

Model selection can be performed using the 'selection' parameter. The available methods ('method') include 'FORWARD', 'BACKWARD', 'STEPWISE', 'LASSO', and 'ELASTICNET'. You can control the selection process with criteria such as 'AIC', 'SBC', or significance levels ('slEntry', 'slStay')....
Read full answer See technical action

Yes, Bayesian analysis is enabled by specifying the 'bayes' parameter. This allows you to define prior distributions for the model parameters using the 'prior' subparameter. The action then uses MCMC methods, such as the No-U-Turn Sampler ('NUTS') or Random Walk Metropolis ('RWM'), to generate poste...
Read full answer See technical action

To save the estimated model, you use the 'store' parameter. You provide a name and optionally a caslib for the output item store. This stored model can then be used by other actions, like 'countreg.countregViewStore', or for scoring new data....
Read full answer See technical action

You can generate an output table with various statistics by using the 'output' parameter and specifying a CAS output table with 'casOut'. Within the 'output' parameter, you can request statistics such as predicted mean ('pred'), probability of the observed count ('prob'), and the linear predictor ('...
Read full answer See technical action

The `countreg.countregViewStore` action displays information about a regression model that has been restored from an item store....
Read full answer See technical action

The required parameters for the `countreg.countregViewStore` action are `instore` and `table`....
Read full answer See technical action

The `instore` parameter specifies the input item store from which a regression model will be restored, allowing various reports about the model to be displayed....
Read full answer See technical action

The `table` parameter specifies the input data table for the action....
Read full answer See technical action

The `viewOptions` parameter allows you to specify which reports to display. Options include `all`, `classVarLevels`, `correlations`, `covariances`, `finalEstimates`, `fitModelSummary`, `initialEstimates`, `minimal` (default), `modelDefinition`, `none`, and `optimizerSettings`....
Read full answer See technical action

You can use the `outputTables` parameter, specifically its `names` subparameter, to specify the list of display tables you want to save as CAS tables....
Read full answer See technical action

If the `display` parameter is omitted, the action creates all possible display tables by default....
Read full answer See technical action

The `phreg.cox` action is used to fit a Cox proportional hazards regression model for survival data....
Read full answer See technical action

To run a basic analysis, you must specify the input data table using the `table` parameter and define the model structure using the `model` parameter, which includes specifying the dependent variable (time-to-event) and explanatory variables....
Read full answer See technical action

Censored data is handled within the `model` parameter. You specify the censoring variable using the `censor` subparameter and the numeric value(s) that indicate a censored observation using the `censVals` subparameter....
Read full answer See technical action

The action supports several model selection methods via the `selection` parameter, including `FORWARD`, `BACKWARD`, `STEPWISE`, and `LASSO`. You can control the selection process with criteria like `slEntry` (significance level for entry) and `slStay` (significance level to stay)....
Read full answer See technical action

Yes, you can perform a stratified analysis by specifying a stratification variable in the `strata` parameter. This fits a separate baseline hazard function for each stratum....
Read full answer See technical action

You can use the `output` parameter to create an output table. You must specify the output table details using the `casOut` subparameter. You can then request various statistics to be included, such as `xBeta` (linear predictor), `survival` (survival probability), `cumHaz` (cumulative hazard), and di...
Read full answer See technical action

The `ties` parameter specifies how to handle tied event times. The available methods are `BRESLOW`, `EFRON`, `EXACT`, and `DISCRETE`. The default is `BRESLOW`....
Read full answer See technical action

Time-dependent covariates can be included by specifying an `entry` variable within the `model` parameter. This allows for modeling covariates whose values change over time....
Read full answer See technical action

The `createBackup` action is used to create a backup, but only if one is not already in progress....
Read full answer See technical action

The `createBackup` action accepts the following parameters:
- `path`: A string that specifies the directory path for the backup location.
- `tables`: A boolean (TRUE/FALSE) that, when set to True, backs up each object as a CAS table instead of persisting the backup to the file system. The default is...
Read full answer See technical action

The `path` parameter is a required string that specifies the directory path where the backup will be stored....
Read full answer See technical action

The `tables` parameter is an optional boolean. If set to TRUE, the backup will store each object as a CAS table. If FALSE or omitted, the backup is persisted to the file system in the specified path....
Read full answer See technical action

The crfScore action scores documents against a pre-existing Conditional Random Fields (CRF) model....
Read full answer See technical action

The required parameters are 'casOut' to specify the output table, 'model' to provide the input model tables, 'table' for the input data, and 'target' to name the predicted variable in the output....
Read full answer See technical action

The 'model' parameter specifies the input modeling tables, which include 'attr' (attributes), 'attrfeature' (attribute-feature mapping), 'feature' (features), 'label' (labels), and 'template' (templates)....
Read full answer See technical action

The output table, which contains the tagged data, is specified using the 'casOut' parameter. You must provide a name for the table and can optionally specify a caslib....
Read full answer See technical action

The 'target' parameter is used to specify the name of the column that will contain the predicted labels (the hidden variable) in the output table....
Read full answer See technical action

The crfTrain action is used to train a Conditional Random Fields (CRF) model for sequence labeling tasks....
Read full answer See technical action

The required parameters for the crfTrain action are 'table' to specify the input data, 'model' to define the output modeling tables, 'target' for the predicted variable, and 'template' to specify the feature extraction template....
Read full answer See technical action

The 'model' parameter specifies the various output tables for the trained model. This includes tables for labels ('label'), attributes ('attr'), features ('feature'), attribute-feature mappings ('attrfeature'), and the template itself ('template')....
Read full answer See technical action

The 'target' parameter specifies the name of the variable in the input table that contains the predicted or hidden labels for the sequence....
Read full answer See technical action

You specify the feature extraction logic using the 'template' parameter, which takes a string containing the textual template that defines how features are to be extracted from the data....
Read full answer See technical action

You can specify the optimization solver using the 'algorithm' subparameter within 'nloOpts'. The available options are: 'LBFGS' (limited-memory Broyden-Fletcher-Goldfarb-Shanno), 'SGD' (stochastic gradient descent), 'ADAM' (adaptive moments variant of SGD), and 'HF' (Hessian free, which is experimen...
Read full answer See technical action

You can apply L1 and L2 regularization using the 'regL1' and 'regL2' parameters, respectively, within the 'optmlOpt' subparameter of 'nloOpts'. Both values must be non-negative....
Read full answer See technical action

When using the 'LBFGS' algorithm, you can specify the line search method via the 'lineSearchMethod' subparameter within 'lbfgsOpt'. The options are 'MORETHUENTE' (default), 'ARMIJO', 'WOLFE', 'STRONGWOLFE', and 'HYBRID'....
Read full answer See technical action

The crossTab action is used to perform one-way or two-way tabulations on a given dataset....
Read full answer See technical action

To perform a two-way tabulation, you must specify both the 'row' parameter for the row variable and the 'col' parameter for the column variable....
Read full answer See technical action

When the 'chiSq' parameter is set to True, the action computes chi-square statistics to test the independence of the row and column variables, including their asymptotic p-values. The default is False....
Read full answer See technical action

You can include missing values in the crosstabulation by setting the 'includeMissing' parameter to True. By default, it is False....
Read full answer See technical action

The 'weight' parameter specifies a numeric variable whose values are used to compute weighted statistics for each cell in the table, as well as for the margins. The specific statistic is determined by the 'aggregator' parameter....
Read full answer See technical action

Set the 'association' parameter to True to compute various measures of association between the row and column variables of the crosstabulation. The default is False....
Read full answer See technical action

The 'aggregator' parameter specifies the type of statistic to compute when a 'weight' variable is used. Options include "SUM", "MEAN", "N" (number of observations), "STD" (standard deviation), and many others....
Read full answer See technical action

The crossValidate action is used to perform cross-validation with specified machine learning actions....
Read full answer See technical action

The crossValidate action supports several model types, including 'BNET' (Bayesian Network Classifier), 'DECISIONTREE', 'FACTMAC' (Factorization Machine), 'FOREST', 'GRADBOOST' (Gradient Boosting Tree), 'NEURALNET', and 'SVM' (Support Vector Machine)....
Read full answer See technical action

The 'kFolds' parameter specifies the number of folds to use for the cross-validation process. The default value is 5, with a minimum required value of 2....
Read full answer See technical action

The 'trainOptions' parameter is a required dictionary that allows you to specify all the necessary parameters for the model training action that will be used during the cross-validation....
Read full answer See technical action

Yes, by default, the folds are evaluated in parallel. You can control this behavior with the 'parallelFolds' parameter (defaulting to True) and specify the number of worker nodes for each subsession with the 'nSubsessionWorkers' parameter....
Read full answer See technical action

You can use the 'seed' parameter to specify a seed for fold sampling, which ensures reproducibility of the cross-validation process. The default value is 0....
Read full answer See technical action

The 'logLevel' parameter controls the verbosity of log messages. It ranges from 0 (no logs) to 3 (most detailed logs, including fold start and completion). The default is 3....
Read full answer See technical action
Copyright © WeAreCAS 2023-2025