Personally identifiable information detection

摘要

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for privacy protection. In one aspect, a method includes accessing personally identifiable information (PII) type definitions that characterize PII types; identifying PII type information included in content of a web page, the PII type information being information matching at least one PII type definition; identifying secondary information included in the content of the web page, the secondary information being information that is predefined as being associated with PII type information; determining a risk score from the PII type information and the secondary information; and classifying the web page as a personal information exposure risk if the risk score meets a confidentiality threshold, wherein the personal information exposure risk is indicative of the web page including personally identifiable information.

主权项

1. A method performed by data processing apparatus, the method comprising:
accessing, by a data processing apparatus, personally identifiable information (PII) type definitions that characterize PII types; identifying, by the data processing apparatus, PII type information included in content of a web page, the PII type information being information matching at least one PII type definition; identifying a sup-portion of content of the web page, the sub-potion of content being content within a window that includes the PII type information and additional content and excluding other content of the web page; identifying, by the data processing apparatus, secondary information included in the sub-portion of content of the web page, the secondary information being content that matches information that is predefined as being associated with PII type information; determining a risk score from the PII type information and the secondary information; and classifying the web page as a personal information exposure risk if the risk score meets a confidentiality threshold, wherein the personal information exposure risk is indicative of the web page including personally identifiable information.