Frequently Asked Questions

550 questions found.

Page 7 / 11

You must use the 'pom' parameter. For each potential outcome, you need to specify the treatment level ('trtLev') and can optionally provide the variable for predicted treatment probability ('trtProb') and the variable for predicted counterfactual outcome ('predOut')....

The type of outcome is specified using the 'type' subparameter within the 'outcomeVar' parameter. It can be 'CONTINUOUS' for real-valued outcomes, 'CATEGORICAL' for modeling the probability of a specific event level, or 'BINOMIAL' for event/trial data....

Yes, by setting the 'inference' parameter to TRUE, the action will compute standard errors and confidence intervals for the potential outcome mean and causal effect estimates. The significance level can be adjusted with the 'alpha' parameter....

You can specify the estimation method using the 'method' parameter. The available options are 'AIPW' (Augmented Inverse Probability Weighting), 'IPW' (Inverse Probability Weighting), 'REGADJ' (Regression Adjustment), and 'TMLE' (Targeted Maximum Likelihood Estimation)....

The outcome variable is specified using the 'outcomeVar' parameter. You must provide its name and can also define its type ('CONTINUOUS', 'CATEGORICAL', 'BINOMIAL'), the event level for categorical outcomes, or the trial variable for binomial outcomes....

The 'pom' (Potential Outcome Means) parameter is required to specify the potential outcomes you want to estimate. For each potential outcome, you must define the treatment level ('trtLev') and can provide the predicted counterfactual outcome ('predOut') and the predicted treatment probability ('trtP...

To compute standard errors and confidence intervals, set the 'inference' parameter to TRUE. The significance level for these intervals can be adjusted using the 'alpha' parameter, which defaults to 0.05....

The 'IPW' (Inverse Probability Weighting) method requires the observed outcome and predicted treatment probabilities. The 'AIPW' (Augmented Inverse Probability Weighting) method is a doubly robust method that also requires predicted counterfactual outcome values in addition to what IPW requires....

The caslibInfo action shows information about caslibs....

When set to True and the 'caslib' parameter is not specified, it shows information only for the active caslib. The default is FALSE....

Use the 'caslib' parameter (alias 'lib') to specify the name of the caslib. If you don't specify it, information for all caslibs is shown....

Set the 'showHidden' parameter to True to include hidden caslibs in the results when requesting information for all caslibs. The default is FALSE....

It specifies the type of caslibs to show. You can choose from 'ALL', 'DNFS', 'ESP', 'LASR', 'PATH', or 'S3'. This parameter is ignored if a specific caslib is named. The default is 'ALL'....

When set to True, the action returns more detailed results. The default is FALSE....

The catTrans action is used to group and encode categorical variables using a variety of unsupervised and supervised grouping techniques....

The catTrans action provides several methods for transformation: 'DTREE' (groups based on a one-level decision tree), 'GROUPRARE' (groups rare levels of the variable, which is an unsupervised technique), 'ONEHOT' (performs one-hot encoding), 'RTREE' (groups based on a one-level regression tree), and...

You can use the 'GROUPRARE' method. The 'preprocessRare' parameter, when set to True, groups rare levels at the start. The rarity is determined by 'rareThreshold' (a frequency count) or 'rareThresholdPercent' (a percentage of total observations). Levels falling below this threshold are combined into...

The supervised techniques require a target variable and include 'DTREE' (decision tree), 'RTREE' (regression tree), and 'WOE' (Weight of Evidence). These methods use the target variable to guide the grouping process....

You can control missing value handling with the 'binMissing' parameter. When set to True, it bins all missing values into a separate bin with an ID of 0. Additionally, the 'missingBinStats' parameter determines if this missing bin should be included when computing evaluation statistics....

The catTrans action is used to group and encode categorical variables using various unsupervised and supervised grouping techniques....

The catTrans action supports several methods specified by the 'method' parameter: 'DTREE' (decision tree grouping), 'GROUPRARE' (grouping rare levels), 'ONEHOT' (one-hot encoding), 'RTREE' (regression tree grouping), and 'WOE' (Weight of Evidence grouping)....

You can use the 'GROUPRARE' method. This method groups rare levels together. You can control this by setting 'preprocessRare' to TRUE and defining the threshold with 'rareThreshold' (for frequency count) or 'rareThresholdPercent' (for percentage)....

One-hot encoding transforms a categorical variable into a set of binary variables (one for each level). To use it, you set the 'method' parameter to 'ONEHOT'. This is also aliased as 'LABEL'....

The supervised techniques are 'DTREE', 'RTREE', and 'WOE'. These methods require specifying one or more target variables using the 'targets' parameter, as the grouping is based on the relationship with these targets....

The 'WOE' (Weight of Evidence) method groups variables by maximizing the Information Value (IV). Key options include 'woeAdjust' to handle bins with zero events or non-events, and 'woeDefinition' which can be set to 'EVENT' (ln(event/non-event)) or 'NONEVENT' (ln(non-event/event))....

Yes, you can generate SAS DATA step scoring code by specifying the 'code' parameter. This will create a file containing the DATA step code that applies the transformations....

The cChart action produces c charts, which are control charts for the number of nonconformities (or defects) per inspection unit....

The `allN` parameter, when set to True, includes all subgroups in the analysis, regardless of whether their sample size equals the nominal sample size specified. The default is False....

You can use the `chartsTable` parameter to specify a CAS output table for the chart summary....

The `exChart` parameter, when set to True, ensures that a control chart is included in the results only when exceptions (such as points outside control limits) are detected. Its default value is False....

Use the `limitN` parameter to specify a nominal sample size for the control limits....

The `primaryTests` parameter is used to request one or more tests for special causes (also known as Western Electric rules) for the primary control chart. These tests help detect non-random patterns in the data....

`test1` checks for one point that falls beyond Zone A, which means it is outside the upper or lower control limits....

`test2` identifies a pattern where nine consecutive points fall on the same side of the center line....

The `sigmas` parameter specifies the width of the control limits as a multiple of the standard error of the subgroup summary statistic. The default value is 3, corresponding to 3-sigma limits....

Yes, by setting the `no3SigmaCheck` parameter to True, you can enable tests for special causes even when the control limits are not the standard three-sigma limits....

The cChart action is part of the Statistical Process Control (SPC) action set and is used to produce c-charts. These charts are used to analyze and monitor the number of nonconformities per inspection unit in a given process....

The 'table' parameter is required. It specifies the input CAS table that contains the process measurement data you want to analyze....

You can use the 'primaryTests' parameter. It's a list where you can enable specific tests. For instance, setting 'test1=True' requests a test for one point beyond Zone A (outside the control limits). Other tests (test2 through test8) detect different non-random patterns....

The 'sigmas' parameter defines the width of the control limits. It is a multiple of the standard error of the subgroup summary statistic. The default value is 3, which creates standard 3-sigma control limits....

Yes, you can save the control limits to an output data table by specifying the 'outLimitsTable' parameter. This creates a new CAS table containing the control limit values....

By default, tests for special causes are not performed when subgroup sample sizes vary. To enable these tests in such cases, you must set the 'testNStd' parameter to True....

The astore.check action, part of the Analytic Store Scoring action set, is used to verify if an ONNX model is valid....

The astore.check action requires a single parameter: 'onnx', which specifies the ONNX model to be checked....

The ONNX model is provided as a binary large object (BLOB) to the 'onnx' parameter....

The CASL syntax is: aStore.check / onnx=; where is your ONNX model....

The checkOutObject action reserves an object and all of its children for update exclusively by the current client session. It can also prevent an object and its parents from being checked out exclusively by another session when using the 'Shared' checkOutType....

The 'checkOutType' parameter accepts two values: "EXCLUSIVE" and "SHARED". "EXCLUSIVE" is the default and ensures only the current session can update the object. "SHARED" prevents other sessions from getting an exclusive lock on the object and its parents, without locking the object itself....

To specify the object, you must use the 'ObjectSelector' parameter. Within this parameter, you define the 'objType' (such as "TABLE", "CASLIB", "COLUMN", "ACTION", or "ACTIONSET") and provide the corresponding identifying parameters like 'caslib', 'table', or 'actionSet'....

The 'checkoutParent' parameter is a boolean that, if set to TRUE, instructs the action to check out the parent object in case the specified object does not exist. The default value is FALSE....