Setting Options

Option Name	Description
Methods
Classification criterion method	specifies the method to use in deriving the classification criterion. If you select Parametric, a parametric method based on the normal distribution within each class is used to derive a linear or quadratic discriminant function. You can also choose from these nonparametric methods: K nearest neighbors. You must specify the value for K. Kernel density estimation. You must specify the value for the radius and the kernel density to estimate the group-specific densities.
Discriminant function	determines whether the pooled or within-group covariance matrix is the basis of the measure of the squared distance. If you select Linear, the task uses the pooled covariance matrix in calculating the (generalized) squared distances. Linear discriminant functions are computed. If you select Quadratic, the task uses the individual within-group covariance matrices in calculating the distances. Quadratic discriminant functions are computed.
Canonical analysis	specifies whether to perform a canonical discriminant analysis. You can also specify the number of canonical variables to compute. This number must be less than or equal to the number of variables.
Validation
Perform cross validation	specifies the cross validation classification of the input data set. When a parametric method is used, each observation in the input data set is classified using a discriminant function. This function is computed from the other observations in the input data set, excluding the observation being classified. You can specify whether to display the cross validation classification results for misclassified observations. When a nonparametric method is used, the covariance matrices used to compute the distances are based on all observations in the data set, excluding the observation being classified. However, the observation being classified is excluded from the nonparametric density estimation or the k nearest neighbors of that observation.
Perform data set validation	specifies the data set that contains the observations to be classified. The names of the quantitative variables in this data set must match the names in the input data set.
Create classification data set	creates an output SAS data set that contains all the data from the Data set to classify data set, plus the posterior probabilities and the class into which each observation is classified. If you select the Canonical analysis check box, the data set also contains new variables with canonical variable scores.
Create classification density data set	creates an output SAS data set that contains all the data from the Data set to classify data set, plus the group-specific density estimates for each observation.
Statistics
Squared Mahalanobis distances	displays the squared Mahalanobis distances between the group means, F statistics, and the corresponding probabilities of greater Mahalanobis squared distances between the group means.
Posterior probability error-rate estimates	displays the posterior probability error-rate estimates of the classification criterion based on the classification results.
Simple descriptive statistics	displays simple descriptive statistics for the total sample and within each class.
Classification results	displays the resubstitution classification results for each observation.
Misclassified observations	displays the resubstitution classification results for misclassified observations only.