Unformatted text preview: Is it the same as the sigma value chosen for newrbe? Part III. Use the method of Laplacian Regularized Least-Squares to solve a semi-supervised classification problem using the same three data sets. For data sets 1 and 3, choose two points at random to be labeled and use all the rest of the data unlabeled. For data sets 2, determine a reasonable number of points to be chosen at random to be able to classify all the data into the two classes given. Explain how you selected parameters for the method (sigma, epsilon, lambda-A, and lambda-I). Discuss your results....
View Full Document
- Spring '08
- Neural Networks, data sets, Laplacian Regularized Least-Squares