. 2021 Jan 15;28(4):744–752. doi: 10.1093/jamia/ocaa327

Table 1.

A summary of the variables used in the framework

Variable	Description
$D = {d_{1}, . ., d_{n}}$	The set of records in the dataset
$F = {f_{1}, \dots, f_{m}}$	The set of attributes in the dataset
$S = {s_{d_{1}}, . . ., s_{d_{n}}}$	The set of data subjects, $s_{d_{i}}$ is the subject associated with a record $d_{i}$
${F'_{i, 1}$ , …, $F'_{i, h_{i}}}$	The set of attributes groups for dataset $s_{d_{i}}$ , $F'_{i, a} \subseteq F$ , $F'_{i, a} \cap F'_{i, b} = \emptyset$ for any $a \neq b$
${P'_{i, 1}$ , …, $P'_{i, h_{i}}}$	The probability that the attacker can obtain each attribute group from ${F'_{i, 1}$ , …, $F'_{i, h_{i}}}$ for an arbitrary subject $s_{d_{i}}$
$p_{m}$	$p_{m} = P (X_{m} = True)$ , the probability that the subject’s presence in the de-identified dataset is disclosed to the attacker.
$p_{c}$	$p_{c} = P (X_{c} = True)$ , the probability that the attacker confirms that a record corresponds to the subject of interest.
$p_{f m}$	$p_{f m} = (X_{l m} = True \| X_{m} = True and$ $X_{u d} = True)$ , the probability that the attacker discovers that the subject is in the dataset given that the subject’s presence in the dataset is known.
$p_{c u}$	$p_{c u} = P (X_{l u} = True\| X_{u p} = True)$ , the probability that the attacker confirms that a subject is unique in the population for a set of attributes given that the data subject is unique.