m-Privacy for Collaborative Data Publishing

Abstract:
we consider the collaborative data publishing problem for anonymizing horizontally
partitioned data at multiple data providers. We consider a new type of insider attack by
colluding data providers who may use their own data records (a subset of the overall data) in
addition to the external background knowledge to infer the data records contributed by other data
providers. he paper addresses this new threat and makes several contributions. !irst" we
introduce the notion of m#privacy" which guarantees that the anonymized data satisfies a given
privacy constraint against any group of up to m colluding data providers. $econd" we present
heuristic algorithms exploiting the e%uivalence group monotonicity of pri#vacy constraints and
adaptive ordering techni%ues for efficiently checking m#privacy given a set of records. !inally"
we present a data provider#aware anonymization algorithm with adaptive m#privacy checking
strategies to ensure high utility and m#privacy of anonymized data with efficiency. &xperiments
on real#life datasets suggest that our approach achieves better or comparable utility and
efficiency than existing and baseline algorithms while providing m#privacy guarantee.
GLOBALSOFT TECHNOLOGIES
IEEE PROJECTS & SOFTWARE DEVELOPMENTS
IEEE FINAL YEAR PROJECTS|IEEE ENGINEERING PROJECTS|IEEE STUDENTS PROJECTS|IEEE
BULK PROJECTS|BE/BTECH/ME/MTECH/MS/MCA PROJECTS|CSE/IT/ECE/EEE PROJECTS
CELL: +91 9!9" #9$"% +91 99&&' #"(% +91 9!9" "(9$% +91 9($1! !$!$1
V)*)+: ,,,-.)/012304546738+*-649 M0)1 +6:)333.)/01*3:546738+*;9:0)1-86:
GLOBALSOFT TECHNOLOGIES
IEEE PROJECTS & SOFTWARE DEVELOPMENTS
IEEE FINAL YEAR PROJECTS|IEEE ENGINEERING PROJECTS|IEEE STUDENTS PROJECTS|IEEE
BULK PROJECTS|BE/BTECH/ME/MTECH/MS/MCA PROJECTS|CSE/IT/ECE/EEE PROJECTS
CELL: +91 9!9" #9$"% +91 99&&' #"(% +91 9!9" "(9$% +91 9($1! !$!$1
V)*)+: ,,,-.)/012304546738+*-649 M0)1 +6:)333.)/01*3:546738+*;9:0)1-86:
Architecture 1:
Architecture 2:
Existing System:
We assume the data providers are semi#honest " commonly used in distributed
computation setting. hey can attempt to infer additional information about data coming from
other providers by analyzing the data received during the anonymization. ' data recipient" e.g.
()" could be an attacker and attempts to infer additional information about the records using the
published data () and some background knowl# edge (*+) such as publicly available external
data.
Propose System:
We consider the collaborative data pub#lishing setting (!igure ,*) with
horizontally partitioned data across multiple data providers" each contributing a subset of records
i. 's a special case" a data provider could be the data owner itself who is contributing its own
records. his is a very common scenario in social networking and recommendation systems. -ur
goal is to publish an anonymized view of the integrated data such that a data recipient including
the data providers will not be able to compromise the privacy of the individual records provided
by other parties.
!oules :
,. (atient .egistration
/. 'ttacks by &xternal 0ata .ecipient 1sing 'nonymized 0ata
2. 'ttacks by 0ata (roviders 1sing 'nonymized 0ata and heir -wn 0ata
3. 0octor 4ogin
5. 'dmin 4ogin

!oules Description
Patient "egistration:
6n this module if a patient have to take treatment"he7she should register
their details like 8ame"'ge"0isease they get affected"&mail etc.hese details are maintained in a
0atabase by the 9ospital management.-nly 0octors can see all their details.(atient can only see
his own record.
#ASED $% &'(S PAPE":
When the data are distributed among multiple data providers or data owners" two main settings
are used for anonymization . -ne approach is for each provider to anonymize the data
independently (anonymize#and#aggregate"!igure ,')" which results in potential loss of
integrated data utility. ' more desirable approach is collaborative data publishing which
anonymizes data from all
providers as if they would come from one source (aggregate#and#anonymize" !igure ,*)" using
either a trusted third#party(() or $ecure :ulti#party ;omputation ($:;) protocols to do
computations .
Attac)s by External Data "ecipient *sing Anonymi+e Data.<
' data recipient" e.g. ()" could be an attacker and attempts to infer additional information about
the records using the published data () and some background knowl# edge (*+) such as
publicly available external data.
Attac)s by Data Proviers *sing Anonymi+e Data an &heir $,n Data:
&ach data provider" such as (, in !igure ,"can also use anonymized data and his own data
(,) to infer additional information about other records. ;ompared to the attack by the external
recipient in the first attack scenario" each provider has additional data knowledge of their own
records" which can help with the attack. his issue can be further worsened when multiple data
providers collude with each other.
-(.*"E 1
-(.*"E:2
Doctor /ogin:
6n this module 0octor can see all the patients details and will get the background
knowledge(*+)"by the chance he will see horizontally partitioned data of distributed data base of
the group of hospitals and can see how many patients are affected without knowing of individual
records of the patients and sensitive information about the individuals.
Amin /ogin:
6n this module 'dmin acts as rusted hird (arty(().9e can see all individual
records and their sensitive information among the overall hospital distributed data
base.'nonymation can be done by this people.9e7$he collected informations from various
hospitals and grouped into each other and make them as an anonymised data.
System Configuration:-
'01 System Configuration:-
Processor - Pentium 2(((
$peed # ,., =9z
.': # /5> :* (min)
9ard 0isk # /) =*
!loppy 0rive # ,.33 :*
+ey *oard # $tandard Windows +eyboard
:ouse # wo or hree *utton :ouse
:onitor # $?='
S01 System Configuration:-
-perating $ystem <Windows@57@A7/)))7B(
'pplication $erver < omcat5.)7>.B
!ront &nd < 9:4" Cava" Csp
$cripts < Cava$cript.
$erver side $cript < Cava $erver (ages.
0atabase < :y s%l
0atabase ;onnectivity < C0*;.
Conclusion:
6n this paper" we considered a new type of potential at#tackers in collaborative data
publishing D a coalition of data providers" called m#adversary. o prevent privacy disclosure by
any m#adversary we showed that guaranteeing m#privacy is enough. We presented heuristic
algorithms exploiting e%uiv#alence group monotonicity of privacy constraints and adaptive
ordering techni%ues for efficiently checking m#privacy. We introduced also a provider#aware
anonymization algorithm with adaptive m#privacy checking strategies to ensure high utility and
m#privacy of anonymized data. -ur experiments confirmed that our approach achieves better or
comparable utility than existing algorithms while ensuring m#privacy efficiently. here are many
remaining research %uestions. 0efining a proper privacy fitness score for different privacy
constraints is one of them. 6t also remains a %uestion to address and model the data knowledge of
data providers when data are distributed in a vertical or ad#hoc fashion. 6t would be also
interesting to verify if our methods can be adapted to other kinds of data such as set#valued data.