Methods and apparatus to anonymize a dataset of spatial data
First Claim
Patent Images
1. A method to anonymize a dataset of spatial data, comprising:
- generating, via a processor, a spatial indexing structure with spatial data;
establishing a height value associated with the spatial indexing structure to generate a plurality of tree nodes, each of the tree nodes associated with a respective spatial data count;
calculating localized noise budget values for respective ones of the tree nodes based on the height value and an overall noise budget;
anonymizing the tree nodes with an anonymization process, the anonymization process using the localized noise budget values for respective ones of the tree nodes;
calculating error metrics for respective ones of the tree nodes;
comparing the error metrics to a threshold; and
if the error metric for one of the tree nodes exceeds the threshold, modifying at least one of the height value or the localized noise budget value for the one of the tree nodes by applying an ordinary least squares estimate.
2 Assignments
0 Petitions
Accused Products
Abstract
Methods and apparatus are disclosed to anonymize a dataset of spatial data. An example method includes generating a spatial indexing structure with spatial data, establishing a height value associated with the spatial indexing structure to generate a plurality of tree nodes, each of the plurality of tree nodes associated with spatial data counts, calculating a localized noise budget value for respective ones of the tree nodes based on the height value and an overall noise budget, and anonymizing the plurality of tree nodes with a anonymization process, the anonymization process using the localized noise budget value for respective ones of the tree nodes.
27 Citations
21 Claims
-
1. A method to anonymize a dataset of spatial data, comprising:
-
generating, via a processor, a spatial indexing structure with spatial data; establishing a height value associated with the spatial indexing structure to generate a plurality of tree nodes, each of the tree nodes associated with a respective spatial data count; calculating localized noise budget values for respective ones of the tree nodes based on the height value and an overall noise budget; anonymizing the tree nodes with an anonymization process, the anonymization process using the localized noise budget values for respective ones of the tree nodes; calculating error metrics for respective ones of the tree nodes; comparing the error metrics to a threshold; and if the error metric for one of the tree nodes exceeds the threshold, modifying at least one of the height value or the localized noise budget value for the one of the tree nodes by applying an ordinary least squares estimate. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10)
-
-
11. An apparatus to anonymize a dataset of spatial data, comprising:
-
a spatial decomposition manager to generate a spatial indexing structure with spatial data, the spatial decomposition manager to establish a height value associated with the spatial indexing structure to generate a plurality of tree nodes, the tree nodes associated with respective spatial data counts; a privacy budget manager to calculate a localized noise budget values for respective ones of the tree nodes based on the height value and an overall noise budget; a noise allocation engine to anonymize the tree nodes with an anonymization process, the anonymization process using respective ones of the localized noise budget values for respective ones of the tree nodes; and an error analyzer to; calculate error metrics for the tree nodes; compare the error metrics to a threshold; and if the error metric for one of the tree nodes exceeds the threshold, modify at least one of the height value or the localized noise budget value for the one of the tree nodes by applying an ordinary least squares estimate. - View Dependent Claims (12, 13, 14, 15)
-
-
16. A tangible machine readable storage device or storage disk comprising instructions that, when executed, cause a machine to perform a method comprising:
-
generating a spatial indexing structure with spatial data; establishing a height value associated with the spatial indexing structure to generate a plurality of tree nodes, the tree nodes associated with respective spatial data counts; calculating a localized noise budget value for respective ones of the tree nodes based on the height value and an overall noise budget; anonymizing the tree nodes with an anonymization process, the anonymization process using the localized noise budget value for respective ones of the tree nodes; calculating error metrics for the tree nodes; comparing the error metrics for the tree nodes to a threshold; and if the error metric for one of the tree nodes exceeds the threshold, modifying at least one of the height value or the localized noise budget value for the one of the tree nodes by applying an ordinary least squares estimate. - View Dependent Claims (17, 18, 19, 20, 21)
-
Specification