Applying Novel QuickLearn Algorithm for Pattern Recognition

In this paper we describe novel QuickLearn (QL) algorithm for pattern classification which has two steps. The mapping level, first step, includes a multi input-single output mapping function (MF) with fixed weights and input data were presented to it during the training phase. It is only a kind of data shifting and scaling before presenting to next level. After shifting and scaling output of first step which is a scalar value to the range [0,T], we can round it to have an integer value (y). The second step, matching level, only includes an array with T cells called affine look-up table (ALT). The training phase for QL includes only one step, no learning cycles. In this single step, the integer value y is used as a reference address to call and upload the label for corresponding input samples in N cells of ALT (copy label from the cell [y-N/2] to the cell [y+ N/2-1]) (data leakages to N-l neighbor cells). In the testing phase, we need only to recall and introduce value of the cell with the index y as the final output (final winner class). The QL is evaluated and compared with existing supervised neural networks on a variety of some well known pattern classification problems. Elapsed time for training a fresh QL classifier is typically only few milliseconds.