Machine Learning with R Cookbook - Second Edition

How it works...

LLE is a nonlinear dimension reduction method, which computes a low dimensional neighborhood, preserving embeddings of high dimensional data. The algorithm of LLE can be illustrated in these steps: first, LLE computes the k-neighbors of each data point. Secondly, it computes a set of weights for each point, which minimizes the residual sum of errors, which can best reconstruct each data point from its neighbors. The residual sum of errors can be described as , where w_ij = 0 if x_j is not one of x_i's k-nearest neighbor, and for each i, . Finally, find the vector, Y, which is best reconstructed by the weight, W. The cost function ...