System and a method for identifying private user information
AssigneeBi Science (2009) Ltd
InventorsAssaf Toval, Kfir Moyal, Nimrod Sandlerman
Life of this patent
Estimating the term…
Abstract
A method, a computer program, or a computerized system for a first software program to remove private user information from a collection of data items communicated by a second software program to a third software program. The first software program may receive the data collection from the second software program, detect in the data collection a first indicator associated with the particular private user information, detect in the data collection the private user information within a proximity threshold associated with the first indicator, remove the private user information from the data collection to form cleaned data collection, and provide the cleaned data collection to the second software program.
Claims8
A method comprising:
- receiving, from at least one client software, raw clickstream data comprising at least one undetermined data item of a plurality of data items, wherein the plurality of data items comprise:
- at least one first data item comprising a personally identifiable information;
- at least one second data item comprising an indicator of the personally identifiable information, wherein the first data item and the second data item are not overlapping;
- and
- at least one third data item comprising a measure of proximity between the first data item and the second data item;
- preparing the raw clickstream data for analysis, wherein the preparing comprises at least one of:
- associating each particular data item of the raw clickstream data with one or more other data items;
- labelling at least one data item of the raw clickstream data for training an artificial intelligence model;
- and
- using the association as a label;
- analyzing the raw clickstream data using the trained artificial intelligence model to determine the at least one undetermined data item;
- creating a new rule for detecting an indicator and detecting a personally identifiable information within the measure of proximity from the indicator;
- and
- distributing the new rule to at least one client software.
A non-transitory computer readable medium storing computer code executable by a processor to perform a method, the method comprising:
- receiving, from at least one client-software, raw clickstream data comprising at least one undetermined data item of a plurality of data items, wherein the plurality of data items comprise:
- at least one first data item comprising a personally identifiable information;
- at least one second data item comprising an indicator of the personally identifiable information, wherein the first data item and the second data item are not overlapping;
- and
- at least one third data item comprising a measure of proximity between the first data item and the second data item;
- preparing the raw clickstream data for analysis, wherein the preparing comprises at least one of:
- associating each particular data item of the raw clickstream data with one or more other data items;
- labelling at least one data item of the raw clickstream data for training an artificial intelligence model;
- and
- using the association as a label;
- analyzing the raw clickstream data using the trained artificial intelligence model to determine the at least one undetermined data item;
- creating a new rule for detecting an indicator and detecting a personally identifiable information within the measure of proximity from the indicator;
- and
- distributing the new rule to at least one client software.
Dependencies are inferred from the claim text.