Min-Increment Bloom Filters for Private Reach and Frequency

The patent introduces encrypted probabilistic data structures as intermediaries between multiple data sources. These structures (such as encrypted Bloom filters, encrypted histograms, or encrypted moment-based representations) allow computations to be performed on encrypted data without exposing the underlying private information. The intermediary encrypted structure enables accurate computation of union, intersection, and frequency while maintaining privacy through cryptographic protection.

The patent transforms the data representation by changing parameters from raw private information to encrypted statistical moments or encrypted probabilistic structure parameters. By working with encrypted sums, encrypted means, or encrypted histogram bins instead of raw data, the system achieves computation accuracy while the cryptographic encoding protects privacy. The parameter transformation allows mathematical operations to be performed on encrypted values.

The patent extracts only the essential statistical features (such as sums, means, variances, or histogram bin counts) from large data sets and transmits these extracted features instead of the complete raw data. By taking out only the necessary computational elements and discarding redundant information, the system reduces data transmission volume significantly while maintaining sufficient accuracy for computing union, intersection, and frequency metrics.

The patent segments large data sets into smaller probabilistic data structures or statistical moment representations that can be processed independently. By dividing the data into manageable encrypted segments (such as multiple encrypted histogram bins or segmented Bloom filters), the system reduces the quantity of data that needs to be transmitted and processed as a single unit, while still achieving accurate aggregate computations.