[OASIS Issue Tracker] (COEL-169) Review draft Spec Vs Anonymisation Decision Making Framework

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

    [ https://issues.oasis-open.org/browse/COEL-169?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=67183#comment-67183 ] 

Joss Langford commented on COEL-169:
------------------------------------

I have previously worked through Mark's book and picked out the following observations:
Other than the use of aggregate results from a suitably large population of atoms, the following techniques could potentially be used on the various fields to provide an anonymous atom data set.
WHEN – perturbation would be appropriate to shift data points it time so that they can be directly linked to other records, e.g. phone records.
WHAT – a k-anonymity approach could systematically raise the reported activities in the dataset to a level in the hierarchy where there are sufficient data points in each class.
WHO - a k-anonymity approach using the information in the Segment Data would create groups of individuals of a sufficient size.
HOW – probably not an issue, but a k-anonymity approach can remove any sparsely populated classes.
WHERE – the easiest route is remove location but either relative distances (between atoms) or rough locations with k-anonymity could be used. Some perturbation may be needed for relative distances to eliminate the potential for specific repeated journeys to be identified.
CONTEXT - probably not an issue, but a k-anonymity approach can remove any sparsely populated classes.
CONSENT – I assume that this would be removed in most cases but, where relevant, certain fields could be used with appropriate k-anonymity.
EXTENSIONS – would need to be looked at on a case-by-case basis but k-anonymity, redaction and perturbation would probably be the best tools.
Overall, when the original atom dataset is large, random sampling would also improve confidence in any disclosure. The use of differential privacy could also allow a more detailed integration (including more information about causality) of a native atoms data set.


> Review draft Spec Vs Anonymisation Decision Making Framework
> ------------------------------------------------------------
>
>                 Key: COEL-169
>                 URL: https://issues.oasis-open.org/browse/COEL-169
>             Project: OASIS Classification of Everyday Living (COEL) TC
>          Issue Type: Improvement
>            Reporter: Matthew Reed
>            Priority: Minor
>              Labels: request_tc_discussion
>
> The UKANON team published a practical Anonymisation Decision Making Handook in 2016. This includes a positive introduction by the UK Information Commisioner: implying that it is the approach the ICO would support.   
> The Handbook is HERE: http://ukanon.net/wp-content/uploads/2015/05/The-Anonymisation-Decision-making-Framework.pdf



--
This message was sent by Atlassian JIRA
(v6.2.2#6258)