1089-49-325 Lori Beth Ziegelmeier* (ziegelme@math.colostate.edu), 101 Weber Building, Fort Collins, CO 80523-1874, and Michael Kirby (kirby@math.colostate.edu) and Chris Peterson (peterson@math.colostate.edu). Sparse Nearest Neighbor Selection for the Locally Linear Embedding Algorithm. Manifold learning techniques such as the Locally Linear Embedding (LLE) algorithm have been proven useful in geometric data analysis and dimensionality reduction. We present modifications to the LLE algorithm that lead to sparse representations in local reconstructions by using a data weighted l-1 norm regularization added to the reconstruction error. This new formulation has proven effective at automatically determining nearest neighbors using sparsity of numerical results. We apply this technique to biological data sets such as gene expression data from the Duke influenza study. (Received February 18, 2013) 1