The estimation of the optimum linear decision function with a sequential random method

This paper is concerned with the categorizing of patterns by means of linear decision functions,
^{1}
as developed by Highleyman [1], [2]. A new procedure for the estimation of the optimum linear decision function is developed. This procedure determines the minimum estimation of the loss function with a sequential random method. An application of the procedure to the same problem as discussed in [1] and [2] yields better results than the procedure developed by Highleyman.