Systems and methods of classifying a subject data item based on a training set of pre-classified data items. A piecewise-linear approximation of a local boundary between different classes of data items is automatically computed. The local boundary is approximated by a neighborhood set of data items selected from the training set that have been pre-classified into different classes and have points similar to a point of the subject data item. A class is automatically assigned to the subject data item in accordance with a side of the local boundary on which the subject data item resides.