| 摘要 |
The invention relates to a method for the automated comparison of two sets of measuring values, wherein the measuring values of the two sets are each associated to a class of a finite number of classes defined by indices such that one frequency distribution is defined for each of the two sets, which indicates for each class a frequency of the measurement values falling within said class, whereupon a distance measurement between said frequency distributions is calculated as a function of the final value of a first auxiliary value, which is denoted, for example, as a match, wherein the first auxiliary value match is calculated by means of an algorithm using two sets of variables, in that, at a given maximum distance dmax = 1, an actual value of a further auxiliary value is defined as m = min(qi', vj') for all integer distances d at 0 = d = dmax, beginning with d = 0 and progressing at greater distances d, for all indices i and j separated from each other by the distance d, wherein m stands for the further auxiliary value, qi' stands for the variables of a first of the two variable sets, and vj ' stands for the variables of the second set of variables, wherein the variables of the two sets of variables are defined at the beginning of the algorithm as qi' = qi, vj' = vj, wherein qi stands for the frequency from one of the first of the two frequency distributions, and vj stands for the frequencies of the second frequency distribution, - whereupon the variables qi' and vj' are redefined by subtracting the actual value of the further auxiliary value m, and the actual value of the further auxiliary value m that is multiplied by a matrix element ai,,j is added to an actual value of the first auxiliary value match originally defined as match = 0, wherein the matrix elements ai,,j form a similarity matrix. The invention further relates to a corresponding device for the automated comparison of two sets of measuring values. |
