SEMISUPLEARN Procedure

The SEMISUPLEARN procedure implements the graph-based semisupervised learning algorithm that relies on label spreading (Zhou et al. 2003) in a distributed framework in SAS Viya. The semisupervised learning algorithm has numerous applications, including web page classification, image recognition, medical imaging, natural language processing, and action recognition.

You can use the SEMISUPLEARN procedure to read and write data in distributed form and to perform search in parallel by making full use of multicore computers or distributed computing environments.

The SEMISUPLEARN procedure operates on a labeled data table and an unlabeled data table. For each observation in the query data table, PROC SEMISUPLEARN returns the predicted labels for the observations in both the unlabeled data table and the labeled data table. The procedure uses the radial basis function (RBF) kernel metric and k-nearest neighbor (KNN) metric to compute distances between the pairs of the observations in the unlabeled and labeled data tables, and it assigns the predicted labels to the unlabeled data in an iterative way, according to the similarity measure. PROC SEMISUPLEARN stores the predicted labels for both the labeled data table and the unlabeled data table in the output data table that you specify in the PROC SEMISUPLEARN statement.