Pattern recognition
A method for pattern recognition performed by a physical computing system includes, with the physical computing system, structuring data as a bi-partite graph, a set of instance nodes within the graph representing instances within the data and a set of property nodes within the graph representing properties of the instances, edges between the instance nodes and the property nodes representing values of the properties, assigning a transition probability function to each of the instance nodes and to each of the property nodes, and applying a random walker to the graph, the random walker utilizing the transition probability functions.
1. A method for pattern recognition performed by a physical computing system, the method comprising:
with said physical computing system, structuring data as a hi-partite graph, a set of instance nodes within said graph representing instances within said data and a set of property nodes within said graph representing properties of said instances, edges between said instance nodes and said property nodes representing values of said properties;
with said physical computing system, assigning a transition probability function to each of said instance nodes and to each of said property nodes; and
with said physical computing system, applying a random walker to said graph, said random walker utilizing said transition probability functions;
in which said transition probability function assigned to each of said property nodes is such that said random walker is more likely to transition between edges closer in value.
2. The method of claim 1 , further comprising, with said physical computing system, determining a pattern in said data by determining which of said edges is more frequently traversed by said random walker.
3. The method of claim 1 , in which a transition probability function assigned to a property node within said set of property nodes comprises one of a continuous transition probability function based on a property associated with that property node and a discrete transition probability function based on a property associated with that property node.
4. The method of claim 1 , in which a transition probability function assigned to a particular instance node within said set of instant nodes comprises a transition probability function configured to make said random walker transition between edges connected to correlated properties.
5. The method of claim 4 , in which said correlated properties are correlated at the value of the arriving edge connected to said particular instance node.
6. The method of claim 4 , in which it is determined how correlated two properties are by:
sorting each edge connected to a first property by value;
sorting each edge connected to a second property based on said sorting of said edges connected to said first property; and
determining if there is a stability of values at any range within said edges connected to said second property.
7. A physical computing system comprising:
a processor; and
a memory communicatively coupled to said processor;
in which said processor is configured to:
structure data as a bi-partite graph, a set of instance nodes within said graph representing instances within said data and a set of property nodes within said graph representing properties of said instances, edges between said instance nodes and said property nodes representing values of said properties;
assign a transition probability function to each of said instance nodes and to each of said property nodes; and
apply a random walker to said graph, said random walker utilizing said transition probability functions;
in which said transition probability function assigned to said property node is such that said random walker is more likely to transition between edges closer in value.
8. The system of claim 7 , in which said processor is further configured to determine a pattern in said data by determining which of said edges is more frequently traversed by said random walker.
9. The system of claim 7 , in which a transition probability function assigned to a property node within said set of property nodes comprises one of: a continuous transition probability function based on a property associated with that property node and a discrete transition probability function based on a property associated with that property node.
10. The system of claim 7 , in which a transition probability function assigned to a particular instance node within said set of instant nodes comprises a transition probability function configured to make said random walker transition between edges connected to correlated properties.
11. The system of claim 10 , in which said correlated properties are correlated at the value of the arriving edge connected to said particular instance node.
12. The system of claim 10 , in which in which to determine how correlated two properties are, said processor is further configured to:
sort each edge connected to a first property by value;
sort each edge connected to a second property based on said sorting of said edges connected to said first property; and
determine if there is a stability of values at any range within said edges connected to said second property.
13. A computer program product for managing providing a layout for graphic objects, said computer program product comprising:
a computer readable storage device having computer readable code embodied therewith, said computer readable program code comprising:
computer readable program code configured to structure data as a bi-partite graph, a set of instance nodes within said graph representing instances within said data and a set of property nodes within said graph representing properties of said instances, edges between said instance nodes and said property nodes representing values of said properties;
computer readable program code configured to assign a transition probability function to each of said instance nodes and to each of said property nodes;
computer readable program code configured to apply a random walker to said graph, said random walker utilizing said transition probability functions; and
computer readable program code configured to determine a pattern in said data by determining which of said edges is more frequently traversed by said random walker;
in which said transition probability function assigned to each of said instance nodes and property nodes is such that said random walker is more likely to transition between edges closer in value.
14. The computer program product of claim 13 , in which said transition probability function assigned to each property node within said set of property nodes comprises one of: a continuous transition probability function based on a property associated with that property node and a discrete transition probability function based on a property associated with that property node.
15. The computer program product of claim 13 , in which a transition probability function assigned to a particular instance node within said set of instant nodes comprises a transition probability function configured to make said random walker transition between edges connected to correlated properties.