Information-theoretic method for clustering and merging imprecise data
View Patent ↗An information-theoretic method clusters and merges bi-variate normal data or ‘error ellipses’ lying in a plane. Two or more error ellipses are clustered and then merged into a single error ellipse if the information lost in the merging process is sufficiently small. This criterion is numerically implemented and tested in a code developed for this purpose.
1. An information theoretic method of clustering and merging bi-variate normal data comprising the steps of:
storing bi-variate normal data in a memory;
processing the data using a processor coupled to the memory to cluster the data into two or more error ellipses; and
further processing the data using the processor to merge the data into a single replacement error ellipse if the information lost in the merging process is sufficiently small.
2. The method according to claim 1 , further comprising:
processing the data to determine a change in entropy associated with merging the ellipses.
3. The method according to claim 2 , wherein the change in entropy is compared with a predetermined threshold to determine whether information lost in the merging process is sufficiently small.
4. The method according to claim 3 , further comprising:
determining the most probable location of a target based on the merging.
5. The method according to claim 1 , wherein the data is one of global positioning system data; geographic information system data, bioinformatics data and pattern recognition data.
6. A computer program product comprising a computer usable medium having computer program logic stored therein, wherein the computer program logic comprises:
clustering logic for causing a computer to cluster data that includes associated error into two or more error ellipses; and
merging logic for causing the computer to merge the data into a single replacement error ellipse if the information lost in the merging process is sufficiently small.
7. The computer program product according to claim 6 , further comprising:
determining logic for causing the computer to determine a change in entropy associated with merging the ellipses.
8. The computer program product according to claim 7 , wherein the change in entropy is compared with a predetermined threshold to determine whether information lost in the merging process is sufficiently small.
9. The computer program product according to claim 8 , further comprising:
determining logic for causing the computer to determine the most probable location of a target based on the merging.
10. The computer program product according to claim 6 , wherein the data is one of global positioning system data; geographic information system data, bioinformatics data and pattern recognition data.