Method and system for determining the geographic location of a network block

摘要

Described herein are a method and a system to assign geographic locations to network blocks. A particular embodiment of the system includes a set of intermediate assignment generators, each intermediate assignment generator being associated with at least one of a plurality of network data sources, each intermediate assignment generator being configured to generate an intermediate assignment for at least one of the plurality of network data sources, a set of classifiers each coupled to at least one of the intermediate assignment generators, each classifier being associated with at least one of the plurality of network data sources, each classifier being configured to generate at least one classification based on at least one of the intermediate assignments and corresponding training data, and an intermediate assignment selector to select a best intermediate assignment based on the classifications generated by the set of classifiers, the best intermediate assignment corresponding to a geographic location of a network block.

主权项

1. A method of assigning a geographic location to a network block comprising:
obtaining geo-location data from a plurality of network data sources; generating, for each data source, intermediate assignments associated with a network block, each intermediate assignment corresponding to geo-location assignment of the network block based on at least one of the network data sources; generating a plurality of feature vectors, each feature vector including a plurality of attributes associated with a different network data source of the plurality of network data sources, a value for a particular attribute of the plurality of attributes representing a degree to which that attribute is present or absent in a corresponding network data source of the plurality of network data sources; classifying and/or regressing each intermediate assignment using the plurality of feature vectors to generate classifications and/or regressions based on training data, the classifications being a mapping from a discrete or continuous feature space to a discrete set of labels, the training data including a set of feature vectors and corresponding desired outputs for each of the feature vectors of the set of feature vectors; and determining a geographic location of the network block based upon at least one of the intermediate assignment classifications and/or regressions.