Privacy preservation for social networks sequential publishing

Safia Bourahla (1, 2) , Maryline Laurent (3, 4, 5) , Yacine Challal (1, 6)

1 LMCS - Laboratoire de Méthodes de Conception de Systèmes
2 Université Blida 2
3 IP Paris - Institut Polytechnique de Paris
4 TSP - RST - Département Réseaux et Services de Télécommunications
5 R3S-SAMOVAR - Réseaux, Systèmes, Services, Sécurité
6 CERIST - Centre de recherche sur l'Information Scientifique et Technique

The proliferation of social networks allowed creating a big quantity of data about users and their relationships. Such data contain much private information. Therefore, anonymization is required before publishing the data for data mining purposes (scientific research, marketing, decision support, etc). Most of the anonymization works in social networks focus on publishing one instance while not considering the need for anonymizing sequential releases. However, many cases show that sequential releases may infer private information even though individual instances are anonymized. This paper studies the privacy issues of sequential releases and proposes a privacy preserving solution for this case. The proposed solution ensures three privacy requirements (users' privacy, groups' privacy and edges' privacy), and it considers the case where many users and groups may share the same profiles. Some experiments over some complex queries show that the utility of the released data is better preserved than other solutions, with regard to the privacy of users, groups and edges.

Mots clés: Privacy preserving; Social networks; Anonymization; Sequential releases; Privacy preserving

Domaines: Informatique [cs]; Cryptographie et sécurité [cs.CR]; Ordinateur et société [cs.CY]; Web

Safia Bourahla, Maryline Laurent, Yacine Challal. Privacy preservation for social networks sequential publishing. Computer Networks, 2020, 170, pp.107106-1:107106-25. ⟨10.1016/j.comnet.2020.107106⟩. ⟨hal-02921464⟩