1: \begin{abstract}

2: A given set of data-points in some feature space may be associated with a \Schrodinger

3: equation whose potential is determined by the data. This is known to lead to good

4: clustering solutions. Here we extend this approach into a full-fledged dynamical scheme

5: using a time-dependent \Schrodinger equation. Moreover, we approximate this Hamiltonian

6: formalism by a truncated calculation within a set of Gaussian wave functions (coherent

7: states) centered around the original points. This allows for analytic evaluation of the

8: time evolution of all such states, opening up the possibility of exploration of

9: relationships among data-points through observation of varying dynamical-distances among

10: points and convergence of points into clusters. This formalism may be further

11: supplemented by preprocessing, such as dimensional reduction through singular value

12: decomposition or feature filtering.

13: \end{abstract}

14: