End while

}

In this stage, information loss of Sk-best is compared with Gs information loss. As the information loss of Sk-best is less than Gs, a tuple with idn is published with Sk-best generalization.

Pid Age University person

Id3 [15–95] University person

Idn + 1 [26–39] University person Idn + 2 [26–39] University person

. . .

Id1 [22–24] Student Id2 [22–24] Student

> . . .

Idn [44–46] Staff

• SetTp = {(<idn,45,academic>, <idn + 1,26,Non academic>,<idn + 2,39,PhD>)}

Table 4 represents Two anonymized university persons.

• SetKc = {(([22–24],university), ([31–39],staff),([44–46],staff))}

• Snew = (<idn,45,academic>,<idn + 2,26,non-academic>)

the tuples in SetTp and insert them into

• Gs = ([26–45],staff)

{

. . .

174 Data Mining

while S!=0 do

For each tuple t do

set Snew.

SetTp.

• Sk-best = ([44–46],staff)

Table 4. Two anonymized university persons.

5.2. Proposed PASS algorithm

Big data Anonymization (S,K,\$)

Read \$ tuples and insert them into

1. Select K-1 unique tuples which are closest to t among

2. Generalize Snew into Gs.
