{"id":7782,"date":"2024-01-10T12:24:36","date_gmt":"2024-01-10T17:24:36","guid":{"rendered":"https:\/\/www.coginov.com\/?p=7782"},"modified":"2024-01-10T19:01:11","modified_gmt":"2024-01-11T00:01:11","slug":"qoreaudit-the-mechanics-of-the-semantic-contextualization","status":"publish","type":"post","link":"https:\/\/www.coginov.com\/en\/qoreaudit-the-mechanics-of-the-semantic-contextualization\/","title":{"rendered":"QoreAudit: The mechanics of the semantic contextualization."},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\n
\n

\n QoreAudit: The mechanics of the semantic contextualization. <\/h2>\n <\/div>\n \t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

By Alkis Papadopoullos, CEO and CSO at Coginov<\/strong><\/p>

As privacy laws are enacted throughout various countries and regions, the ability to uncover relevant tags indicative of personal or sensitive data (often referred to as PII \u2013 personal identifiable information \u2013 and SII \u2013 sensitive identifiable information) is becoming more and more important. Achieving this requires tools and algorithms to be able to put text-based keywords and expressions in context in order to minimize false positive or false negative hits when discovering PII or SII data.\u00a0 Typical types of said data can be person or organization names, phone numbers, addresses, financial information or records, citizen identification documents, etc.<\/p>

To address this need, we present some ideas around the mechanics of semantic contextualization. Platforms that attempt to discover and extract such information have a set of very tangible expectations such as obtaining a list of possible PII or SII candidates for review and subsequent calls to action regarding disposal, storage, or protection of this data. However, this proves very difficult to accomplish if analysis is based solely on brute force extraction without any context. For example, is what seems to be a person\u2019s name, actually part of street name and hence part of an address? Is a sixteen-digit numeral actually a credit card number? Etc.<\/p>

Machine learning based semantic analysis can help to achieve these goals, primarily because it involves associating to each potential sensitive data piece, the meaning of that data; it is thus equivalent to extracting and storing concepts rather than keywords. By identifying concepts along with named entities (PII and SII data) it is possible to achieve three important goals that are the cornerstones of reliably identifying personal information:<\/p>