Fast Deterministic Approximation of Symmetric Indefinite Kernel Matrices with High Dimensional Datasets