When 'promote' is set to TRUE, the loaded table is given a global scope, making it accessible to other CAS sessions, provided the target caslib is also global....
The 'importOptions' parameter specifies the settings for reading a table from a data source, including the 'fileType' (like CSV, EXCEL, PARQUET) and other file-type-specific parameters....
The loadTableFromDisk action loads a sentiment analysis model (SAM), a category model (MCO), or a concept model (LI) binary file from the client machine and creates a CAS table to contain that binary file....
The 'level' parameter can be set to one of the following string values: "ALL", "DEBUG", "ERROR", "FATAL", "INFO", "TRACE", or "WARN". It can also be set to "NULL" to inherit the level from the parent logger, or "OFF" to disable the logger....
When 'newSessions' is set to TRUE, it specifies that the logging level change will apply not only to the current session but also to any new sessions that are created. Its default value is FALSE....
The 'logistic' action, part of the Regression Action Set, is used to fit various types of logistic regression models for binary, binomial, and multinomial (both nominal and ordinal) response variables....
You define the model using the 'model' parameter. This includes specifying the dependent variable(s) with the 'depVars' subparameter and the explanatory effects using the 'effects' subparameter. You can also define the response distribution (e.g., 'BINOMIAL', 'MULTINOMIAL') and the link function (e....
The 'logistic' action supports several automated model selection methods through the 'selection' parameter. These include 'FORWARD' (forward selection), 'BACKWARD' (backward elimination), 'STEPWISE' (stepwise regression), 'LASSO', and 'ELASTICNET'. You can control the process with criteria like 'AIC...
Yes, you can. The 'store' parameter saves the fitted model to a CAS table as a binary large object (BLOB), which can be restored later for scoring or further analysis. Additionally, the 'output' parameter can create a new CAS table containing observation-wise statistics like predicted probabilities,...
Classification variables are specified using the 'class' parameter. This allows you to define how these variables are parameterized in the model, including setting the reference level ('ref'), the ordering of levels ('order'), and the parameterization style ('param', e.g., 'EFFECT' or 'GLM')....
Yes, the 'logistic' action supports repeated measures analysis through the 'repeated' parameter. You must specify the 'subject' and can define the correlation structure of the repeated measurements using the 'corrtype' subparameter (e.g., 'AR' for autoregressive, 'EXCH' for exchangeable)....
The logisticAssociation action computes indices of rank correlation between predicted probabilities and observed responses, which are used for assessing the predictive ability of a model....
The action requires a model restored from a binary large object (BLOB) specified via the 'restore' parameter, and an input data table specified via the 'table' parameter to be scored and assessed....
You can create a classification table by setting the 'ctable' parameter to TRUE. You can also specify the classification cutpoints using the 'cutpt' parameter....
The association table is created by default. You can ensure its creation by setting the 'association' parameter to TRUE, which computes various rank correlation indices between predicted probabilities and observed responses....
When the 'fitData' parameter is set to TRUE, it specifies that the data being scored were also used to fit the model being assessed. The default value is FALSE....
The 'restore' parameter is a required parameter that specifies the input data to restore a regression model from a binary large object (BLOB), which is necessary for the action to generate scoring code....
Use the 'casOut' parameter to specify the settings for the output table where the generated DATA step code will be stored. You need to provide a name for the table and optionally the caslib....
When the 'pCatAll' parameter is set to TRUE, the generated scoring code will compute the probabilities for all levels of the response variable. By default, this is set to FALSE....
The 'restore' parameter is mandatory. It specifies the input CAS table that contains the saved state (as a binary large object or BLOB) of a logistic regression model, which is required to perform the lack-of-fit test....
You can define the groups by using either the 'nGroups' parameter to specify the number of equal-sized groups based on predicted probabilities, or the 'cutpt' parameter to provide specific cutpoints for creating the partitions....
When set to True, the 'powerAdj' parameter adjusts the number of groups to ensure the Hosmer and Lemeshow test maintains adequate statistical power, which can be useful when the number of observations is low....
The degrees of freedom (df) can be controlled using two parameters. The 'df' parameter directly sets the degrees of freedom, while the 'dfReduce' parameter specifies a value to subtract from the default degrees of freedom, which is typically the number of groups minus 2....
The action requires two main input tables: the 'restore' table, which contains the fitted model information, and the 'table' parameter, which specifies the data table to be used for the lack-of-fit computation....
The 'restore' parameter is used to load a pre-existing regression model from a CAS table (a binary large object or BLOB) to compute the odds ratios....
The logisticScore action is used to create a new table on the server that contains the results from scoring observations using a previously fitted logistic regression model....
You must specify two required parameters: 'table', which defines the input data table to be scored, and 'restore', which specifies the item store containing the fitted model to use for scoring....
You must use the 'casOut' parameter. This is a required parameter where you specify the name and caslib for the output table that will store the scoring results....
Use the 'copyVars' parameter. You can provide a list of specific variable names to copy, or use 'ALL' to copy all variables, or 'ALL_NUMERIC' to copy only the numeric variables....
The 'fitData' parameter, when set to True, indicates that the data being scored is the same data that was used to train the model. This can affect how certain statistics are calculated. The default value is False....
To get the predicted mean probability, use the 'pred' parameter to name the output variable. For confidence intervals on this mean, use the 'lclm' and 'uclm' parameters to name the variables for the lower and upper bounds, respectively. The confidence level is controlled by the 'alpha' parameter, wh...
'pred' (or its alias 'p') is used for the predicted mean value or probability. 'ipred' is used to get the individual predicted value specifically for models with a cumulative link function....
A Type 3 test, as performed by the `logisticType3` action, is a statistical test to evaluate the null hypothesis that all parameters associated with a specific effect in the model are simultaneously equal to zero....
The `restore` parameter is required. It is used to specify the input table that contains the regression model information, restored from a binary large object (BLOB)....
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. WeAreCAS is an independent community site and is not affiliated with SAS Institute Inc.
This site uses technical and analytical cookies to improve your experience.
Read more.