Method and apparatus for ordering electronic data

Method of automatically ordering a plurality of sets of electronic data by means of a data processing unit, comprising the following steps to be performed by said data processing unit:

• at least for a selected group of data sets, determining the distance D between any two data sets, said distance being defined as a function of a pair of two data sets, rendering a numerical value, said function having a first value D0 defined for the case of a pair of identical data sets, the difference of the distance D of any pair to said value D0 being defined to be either greater than or equal zero for all pairs, D−D0≧0, or less than or equal zero for all pairs, D−D0≦0,
• determining the density of distance values over the range of determined distance values,
• determining one or more limiting values, at least some of the limiting values defining an upper boundary of a peak in said density of distance values, respectively, if said difference is defined to be D−D0≧0 for all pairs, and at least some of the limiting values defining a lower boundary of a peak, respectively, if said difference is defined to be D−D0≦0, said limiting values forming an increasing series in case of a plurality of limiting values,
• creating correlation data correlating each data set to a cluster in a hierarchy of clusters, the number of cluster levels in said hierarchy corresponding to the number of limiting values, wherein,
• if said difference is defined to be D−D0≧0 for all pairs,
• the data sets contained in each first level cluster in said hierarchy are related to one another in that for each data set the minimum pairwise distance to other data sets in said cluster is less than the lowest limiting value,
• each higher order cluster in said hierarchy comprises data sets of a group of one or more clusters of lower levels, wherein, if said group comprises more than one cluster, each cluster in this group forms a pair with another cluster in this group for which pair there is at least one data set of one cluster of said pair having a distance from a data set of the other cluster of said pair which is less than that limiting value that is the next higher one in said increasing series of limiting values to that limiting value defining clusters at the next lower level,
• and, if said difference is defined to be D−D0≦0 for all pairs,
• the data sets contained in each first level cluster in said hierarchy are related to one another in that for each data set the maximum pairwise distance to other data sets in said cluster is greater than the highest limiting value,
• each higher order cluster in said hierarchy comprises data sets of a group of one or more clusters of lower levels, wherein, if said group comprises more than one cluster, each cluster in this group forms a pair with another cluster in this group for which pair there is at least one data set of one cluster of said pair having a distance from a data set of the other cluster of said pair, which is greater than that limiting value that is the next lower one in said increasing series of limiting values to that limiting value defining clusters at the next lower level.