1 / 37

L- diversity for Privacy-Preserving Traffic Padding in Web- Based Applications

L- diversity for Privacy-Preserving Traffic Padding in Web- Based Applications. DEMO PRESENTATION. L-DIVERSITY FOR PPTD in WEBAPP. Before -diversity. After -diversity. L-DIVERSITY FOR PPTD in WEBAPP (CNTD). Table T . L-DIVERSITY FOR PPTD in WEBAPP (CNTD). Table T .

ince
Download Presentation

L- diversity for Privacy-Preserving Traffic Padding in Web- Based Applications

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. L-diversityfor Privacy-PreservingTrafficPadding in Web-Based Applications DEMO PRESENTATION

  2. L-DIVERSITY FOR PPTD in WEBAPP Before -diversity After -diversity

  3. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) Table T

  4. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) Table T

  5. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) Let r = Index.first() and form a group Table T

  6. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) (2) Compute padding_cost_matrix(T|r) 8 7 1 9 5 2 3 6 4 8 Table T

  7. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) (3) Find the padding priority Q of 8 w.r.t the ungrouped records of T 8 7 1 9 5 2 3 6 4 8 Table T 3  1 Q = 9  7 2 5  6  4

  8. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round (4) Repeat until equivalence class satisfies l-diversity { 3  1 Q = 9  7 2 5  6  4 • diversity_first = diversity_basis() diversity_basis () = = = 1 diversity_basis () = = 1 (is not l-diverse) Table T Remaining records RR= {7,1,9,5,2,3,6,4} diversity_basis (RR) = = = 0.19 ( thereis l-diversity)

  9. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round (4) Repeat until equivalence class satisfies l-diversity { v 3  1 Q = 9  7 2 5  6  4 diversity_basis () = • pop the head vector v from Q Table T

  10. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round (4) Repeat until equivalence class satisfies l-diversity { v 3  1 Q = 9  7 2 5  6  4 diversity_first = • temp = + v temp = {8, 9} • diversity_second = diversity_basis(temp) diversity_second = = 0.81 Table T

  11. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round (4) Repeat until equivalence class satisfies l-diversity { v 3  1 Q = 9  7 2 5  6  4 diversity_first = temp = {8, 9} diversity_second = 0.81 • if (diversity_second < diversity_first) • { = + v • Mark v with assigned sign } diversity_basis () = = 0.81 (is not l-diverse) Table T

  12. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round (4) Repeat until equivalence class satisfies l-diversity { v 3  1 Q = 7 2 5  6  4 diversity_first = temp = {8, 9,7} diversity_second = 0.47 diversity_second > diversity_first } diversity_basis () = = 0.47 (is l-diverse Table T

  13. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 1st round done (5) Copy to T’ done (6) Mark r and ’s elemetnswith assigned sign done (7) Mark ’s elements in T’ with r (8) Remove ’s elements from Index } 8 8 8 Table T Table T' 9 8 7 Index = 1 5 2 3 6 4

  14. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) done (1)Let r = Index.first() and form a group done 2nd round (2) Compute padding_cost_matrix(T|r) (3)Find the padding priority Q of 1w.r.t the ungrouped records of T done Q = 3 2 5  6  4 (4) Repeat until equivalence class satisfies l-diversity { . . . . . . . . . . } Table T diversity_basis () = = 0.41 (is l-diverse) Index = 1 5 2 3 6 4

  15. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 2nd round done (5) Copy to T’ done (6) Mark r and ‘s elements with assigned sign done (7) Mark ’s elements in T’ with r (8) Remove ’s elements from Index } done done done 20 100 1 1 3 40 1 120 Table T 36 2 130 1 2 Index = 5 6 4 1 3 Table T'

  16. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) done (1)Let r = Index.first() and form a group done 3rd round (2) Compute padding_cost_matrix(T|r) done (3)Find the padding priority Q of 1w.r.t the ungrouped records of T done Q = 6  4 done (4) Repeat until equivalence class satisfies l-diversity { done . . . . . . . . . . } Table T diversity_basis () = = 0.43 (is l-diverse) Index = 5 6 4

  17. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) 3rdround done done done done done done done done done 20 135 5 1 3 40 5 140 Table T 36 2 NULL 150 5 6 Index = 5 4 Table T'

  18. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) done done done done done done done done done done 5 25 135 5 6 42 140 5 Table T 4 50 150 5 Table T'

  19. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) done done done done done done done done done done 5 25 135 5 6 42 140 5 Table T 4 50 150 5 Table T'

  20. L-DIVERSITY FOR PPTD in WEBAPP (CNTD) done done done 90 done 90 90 done 130 done done done 130 done 130 done 150 150 Table T 150 Table T'

  21. L-DIVERSITY FOR PPTD in WEBAPP (AXIOM [Minimum Padding Cost]) T is l-diverse (diversity_basis (T) = 0.19 ) but,.. Total Padding cost = 95 Total Padding cost = 335 pc = 30 pc = 335 pc = 40 pc = 25 Table T Table T' Adopt the algorithm is more economical even if the dataset is initially diverse pc = padding cost

  22. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) Now suppose that after the 2nd round we had : done done done done done done Table T' diversity_basis () = = 0.52 (is not l-diverse) Table T So steps (5), (6), (7) and (8) can’t be applied

  23. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) done else remove Index.first() from Index done done done done done Table T 6 4 Index = 5

  24. L-DIVERSITY FOR PPTD in WEBAPP (ASSIGN( ) function) done (1)Let r =Index.first() and form a group done 4th round (2) Compute padding_cost_matrix(T|r) done (3)Find the padding priority Q of 6w.r.t the ungrouped records of T done Q = 4 done (4) Repeat until equivalence class satisfies l-diversity { done . . . . . . . . . . } Table T diversity_basis () = = 0.64 (is not l-diverse) 4 Index = 6

  25. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) done 4th round else remove Index.first() from Index done done done done done Table T 4 Index = 6

  26. L-DIVERSITY FOR PPTD in WEBAPP (ASSIGN( ) function) done (1)Let r = Index.first() and form a group done 5th round done (2)Find the padding priority Q of 4 w.r.t the ungrouped records of T done Q = NULL done (3) Repeat until equivalence class satisfies l-diversity { done . . . . . . . . . . } Table T diversity_basis () = = 1 (is not l-diversified) 4 Index =

  27. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) done 5th round else remove Index.first() from Index done done done done done Table T NULL Index = 4

  28. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • = { remaining unassigned records of T} done done done done done done Table T

  29. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • for each record v in{ record 5 • for each group in T’{ - group 8 : c_ diversity_basis (5 | 8) = 0.36 - group 1 : c_ diversity_basis (5 | 1) = 0.36 • } • Compute min_pc(Candidate | 5) min_pc(Candidate | 5 ) = {1} • Find the perfect candidate Table T‘ The perfect candidate for 5 is group 1

  30. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • Add 5 to group 1 done • Mark v in T with assigned 5 15 135 1 Table T‘

  31. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • for each record v in{ record 6 • for each group in T’{ - group 8 : c_ diversity_basis (6 | 8) = 0.32 - group 1 : c_ diversity_basis (6 | 1) = 0.30 • } • Compute min_pc(Candidate | 6) min_pc(Candidate | 6 ) = {1} • Find the perfect candidate The perfect candidate for 6 is group 1 Table T‘

  32. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • Add 6 to group 1 done • Mark v in T with assigned 6 22 140 1

  33. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • for each record v in{ record 4 • for each group in T’{ - group 8 : c_ diversity_basis (4 | 8) = 0.46 - group 1 : c_ diversity_basis (4 | 1) = 0.23 • } • Compute min_pc(Candidate | 4) min_pc(Candidate | 4 ) = {1} • Find the perfect candidate The perfect candidate for 4 is group 1

  34. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) • Add 6 to group 1 done • Mark v in T with assigned 4 40 150 1

  35. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) done 4 40 150 1

  36. L-DIVERSITY FOR PPTD in WEBAPP(ASSIGN( ) function) done Total padding cost = 155 < 355 90 90 90 150 150 150 150 150 4 40 150 1

  37. References : - Privacy-Preserving Traffic Padding in Web-Based Applications Wen Ming Liu, Lingyu Wang, Pengsu Cheng and MouradDebbabi - An Improved V-MDAV Algorithm for l-Diversity Han Jian-min, Cen Ting-ting and Yu Hui-gun - A Rough Set Based Efficient l-diversity Algorithm B. K. Tripathy1, G. K. Panda2* and K. Kumaran3; • Thanks for visiting the demo • Feedbacks will be welcome

More Related