Structured relevance—a mechanism to reveal why data is related

• US 9,245,056 B2
• Filed: 02/23/2011
• Issued: 01/26/2016
• Est. Priority Date: 02/23/2011
• Status: Expired due to Fees

• Current Assignee: Micro Focus Software, Inc. (Open Text Corporation)
• Original Assignee: Novell Incorporated (Open Text Corporation)
• Inventors: Isaacson, Scott A., Kimball, Kirk R.
• Primary Examiner(s): Pulliam, Christyann
• Assistant Examiner(s): Mahmood, Rezwanul
• Application Number:
  US13/033,509
• Publication Number:
  US 20120215770A1
• Time in Patent Office: 1,798 Days
• Field of Search: 707/728, 707/748, 707/758
• US Class Current: 1/1
• CPC Class Codes: G06F 16/9038 Presentation of query results

Abstract

A machine receives a group of members of a data set. The machine identifies key symbols from the members of the group or the data set. The machine then calculates, for each key symbol, a weighted magnitude for the key symbol in the group. The machine can then sort the key symbols according to their weighted magnitudes, and filter out common key symbols. The uncommon key symbols, as sorted according to their weighted magnitudes, can form a name for the group.

Claims

1. An apparatus, comprising:

a computer;

an input port to receive a group including members of a data set;

a key symbol identifier to identify a key symbol for the group;

a weighted magnitude calculator to calculate a weighted magnitude for each key symbol based on a number of occurrences for each key symbol in the members divided by a total number of members for the group to form a quotient that is multiplied by a maximum value for that key symbol; and

a sorter to sort the key symbols based on the weighted magnitude for each key symbol;

wherein the weighted magnitudes of the sorted key symbols are used to represent relationships for each member of the group relative to remaining ones of the members, and wherein the group includes a group strength, the group strength is a combination of;

a group order for the group, a group weight for the group, and a group distance for the group, and wherein the combination is calculated to ensure that the group order is more significant than the group weight and the group weight is more significant than the group distance, and wherein the group weight is a maximum number of hops that must be made from any member within the group to reach a core of the data set and the core is a particular set of members of the data set such that each member is the nearest neighboring member of each other member, and wherein the group order is determined by a total number of the members within the group, and wherein the group distance is a calculated average distance for each member of the group to the nearest neighboring member for that member, and wherein the group strength is compared against other calculated group strengths for other groups to determine a strength between the group and the other groups relative to one another.

2. The apparatus according to claim 1, wherein each of the key symbols are operative to identify remaining ones of the key symbols in the members.

3. The apparatus according to claim 1, wherein the sorter includes a common key symbol filter to filter out a common key symbol from each of the symbols, producing an uncommon key symbol.

4. The apparatus according to claim 3, wherein the uncommon key symbol forms a name for the group.

5. The apparatus according to claim 1, wherein the weighted magnitude calculator includes:

a frequency calculator to calculate a frequency of each key symbol relative to the remaining ones of the key symbols in each of the members of the group;

a magnitude calculator to calculate a magnitude of each key symbol from each of the members of the group;

a variation calculator to calculate a variation for each key symbol relative to the frequency for each of the remaining ones of the key symbols in each of the remaining ones of the members of the; and

a product calculator to calculate a weighted magnitude as a product of the magnitude and the variation for each of the members of the group.

6. The apparatus according to claim 5, wherein:

the magnitude calculator includes a maximum value calculator to calculate a the maximum value for each of the key symbols in each of the members of the group; and

the variation calculator includes an inverse variation calculator to subtract the variation of each key symbol from a summed maximum value calculated by summing the maximum values for each of the key symbols in the group.

7. A method, comprising:

using a processor, identifying a group including members of a data set;

identifying key symbols in the group, the group including at least one member of the data set including other members;

determining a weighted magnitude for each key symbol;

sorting the key symbols based on a weighted magnitude for each of the key symbols, the sorted key symbols; and

defining a group strength for the group as a combination of;

a group order for the group, a group weight for the group, and a group distance for the group, wherein the group order is a total number of members present within the group, the group weight is a maximum number of hops that must be made from any member within the group to reach a core of the data set and the core is a particular set of members of the data set such that each member is the nearest neighboring member of each other member, the group distance is a calculated average distance for each member of the group to the nearest neighboring member for that member, and wherein the combination is calculated to ensure that the group order is more significant than the group weight and the group weight is more significant than the group distance and comparing the group strength against other calculated strengths for other groups for determining a strength between the group and the other groups;

wherein the weighted magnitudes of the sorted key symbols represent relationships for each member of the group relative to remaining ones of the members of the group.

8. The method according to claim 7, wherein identifying key symbols further includes identifying each of the key symbols in each of the members.

9. The method according to claim 7, wherein sorting the key symbols further includes filtering common key symbols out from the key symbols, producing uncommon key symbols.

10. The method according to claim 9, wherein sorting the key symbols includes further assigning the uncommon key symbols as names for the group.

11. The method according to claim 7, wherein determining the weighted magnitude for each key symbol further includes:

for each member of the group, determining a frequency of for a particular key symbol of that member the key symbol relative to remaining ones of that member'"'"'s key symbols;

determining a magnitude of the key symbols occurring within each of the members based on a number of occurrences for each key symbol in the members divided by a total number of members for the group to form a quotient that is multiplied by a maximum value for that key symbol;

determining a variation of the key symbols for the members relative to the frequency of each the members and that member'"'"'s key symbols; and

determining the weighted magnitude for each key symbol as a product of that key symbol'"'"'s magnitude and that key symbol'"'"'s variation.

12. The method according to claim 11, wherein determining the variation of the key symbols further includes:

determining a particular maximum value for each of key symbols in each of the members; and

subtracting the variation of each of the key symbols from the particular maximum value for the key symbols for the group, the particular maximum value calculated by summing each member'"'"'s key values with the remaining ones of the member'"'"'s key values.

13. An article comprising a non-transitory storage medium, said non-transitory storage medium having stored thereon instructions, that, when executed by a machine, result in:

identifying a group including a particular member of a data set having multiple other members;

identifying key symbols in the group;

determining a weighted magnitude for each key symbol based at least in part on a number of occurrences for each key symbol in the members divided by a total number of members for the group to form a quotient that is multiplied by a maximum value for that key symbol;

sorting the key symbols according to weighted magnitudes for each of the key symbols; and

defining a group strength for the group as a combination of;

a group order for the group, a group weight for the group, and a group distance for the group, wherein the group order is a total number of members present within the group, the group weight is a maximum number of hops that must be made from any member within the group to reach a core of the data set and the core is a particular set of members of the data set such that each member is the nearest neighboring member of each other member, the group distance is a calculated average distance for each member of the group to the nearest neighboring member for that member, and wherein the combination is calculated to ensure that the group order is more significant than the group'"'"'s weight and the group'"'"'s weight is more significant than the group'"'"'s distance, and comparing the group strength against other calculated strengths for other groups for determining a strength between the group and the other groups;

wherein the weighted magnitudes of the sorted key symbols represent relationships for each member of the group relative to remaining ones of the members of the group.

14. The apparatus according to claim 13, wherein identifying the key symbols includes identifying each of key symbols each of the members of the data set.

15. The article according to claim 13, wherein sorting the key symbols includes filtering common key symbols out from the, producing a plurality of uncommon key symbols.

16. The article according to claim 15, wherein sorting the key symbols includes further assigning the uncommon key symbols as names for the group.

17. The article according to claim 13, wherein determining the weighted magnitudes further includes:

for each member of the group, determining a frequency of the key symbol for the member relative to remaining ones of the key symbols in that member of the group;

determining a magnitude for each key symbol from each of the members of the group;

determining a variation of each key symbol relative to the frequency of each other key symbol for each member of the group; and

determining the weighted magnitude for each key symbol as a product of the magnitude for that key symbol and the variation for that key symbol.

18. The article according to claim 17, wherein determining the variations further includes:

determining a maximum value for each of the key symbols in each of the members of the group; and

subtracting the variation of the each of the key symbols from a maximum value for remaining ones of the key symbols for the group, the maximum value calculated by summing each member'"'"'s key values with the remaining ones of the member'"'"'s key values.