Confidence ratio affinity propagation in ensemble selection of neural network classifiers for distributed privacy-preserving data mining

Kokkinos, Yiannis; Margaritis, Konstantinos G.

doi:10.1016/j.neucom.2014.07.065

Cited by 19 publications

(5 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The sanitizing key, A 2 , is reconstructed by exploiting Eqs. (11), (10), and (1) and the proposed OI-CSA update. It is deployed to generate O′ by which the lossless restoring could be carried out by Eq.…”

Section: Restoration Processmentioning

confidence: 99%

“…A major concern in DM [5,16] is the process of finding out frequent item sets, and subsequently association rules [21] are frequently exploited in numerous areas. The majority of the PPDM [9,11] schemes exploit a transformation that minimizes the convenience of the underlying data when it is applied to data-mining algorithms. Anyhow, these schemes are not capable of adjusting the accuracy and privacy efficiently: "when one is preserved, the other appears to suffer" [10,28].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Opposition Intensity-Based Cuckoo Search Algorithm for Data Privacy Preservation

Shailaja

Rao

2019

Journal of Intelligent Systems

View full text Add to dashboard Cite

Abstract Privacy-preserving data mining (PPDM) is a novel approach that has emerged in the market to take care of privacy issues. The intention of PPDM is to build up data-mining techniques without raising the risk of mishandling of the data exploited to generate those schemes. The conventional works include numerous techniques, most of which employ some form of transformation on the original data to guarantee privacy preservation. However, these schemes are quite multifaceted and memory intensive, thus leading to restricted exploitation of these methods. Hence, this paper intends to develop a novel PPDM technique, which involves two phases, namely, data sanitization and data restoration. Initially, the association rules are extracted from the database before proceeding with the two phases. In both the sanitization and restoration processes, key extraction plays a major role, which is selected optimally using Opposition Intensity-based Cuckoo Search Algorithm, which is the modified format of Cuckoo Search Algorithm. Here, four research issues, such as hiding failure rate, information preservation rate, and false rule generation, and degree of modification are minimized using the adopted sanitization and restoration processes.

show abstract

Section: Restoration Processmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Opposition Intensity-Based Cuckoo Search Algorithm for Data Privacy Preservation

Shailaja

Rao

2019

Journal of Intelligent Systems

View full text Add to dashboard Cite

show abstract

“…The implemented algorithm promises improved scalability than the most recent ones regarding databases with the features of steadily increasing transactions and attributes. Kokkinos [30] proposed a security-preserving mining of data in huge decentralized positions of data that can construct numerous neural networks for group formation. Experimental model on comparisons and benchmark datasets with other diversity dependent process and various conventional techniques were assuring.…”

Section: A State-of-the-art Of Contributionsmentioning

confidence: 99%

Route Map of Privacy Preservation to IOT

Kathavate¹,

Amudhavel²

2018

IJET

View full text Add to dashboard Cite

Data anonymization is the main feature of privacy preservation, and it assists in eradicating the privacy hazard in data preparation in various applications including IoT. Pseudonymity and Anonymization are two significant security factors that were adopted when sensitive data are shared. In medical field, data is usually distributed horizontally with diverse regions carrying a similar set of characteristics for various anonymization techniques.

show abstract

“…Yiannis Kokkinos et al [39] discussed DDM privacy-preserving for ensemble technique. Neural classification is selected by confidence ratio affinity propagation by privacy computing.…”

Section: Papyrusmentioning

confidence: 99%

Approaches and Techniques of Distributed Data Mining : A Comprehensive Study

Urmela¹,

Dr.M.²

2017

IJET

View full text Add to dashboard Cite

-Distributed Data Mining (DDM) has become one of the promising areas of Data Mining (DM). DDM evolved from DM from the urge to mine data from distributed sites. DM paved way for increased computational cost and privacy due to centralized data mining, whereas DDM paved way for decrease in computational cost as well as enhanced data privacy by distributing resources across distributed sites. Mining techniques framed for DM can't be applied for DDM since mining DDM follows a different strategy compared to DM. DDM includes classifier based, agent based and privacy preserving based approaches. In this paper, DDM approaches and techniques is studied in detail. Keyword -Distributed Data Mining, distributed sites, computation cost, classifier approach, agent based, privacy-preserving I. INTRODUCTION Data Mining (DM) is the process of extracting useful information from datasets using DM techniques, namely pattern matching, clustering, rule association, regression, etc. The progressive growth of information technology has paved way to further explore Distributed/Collective Data Mining, Spatial and Geographic Data Mining, Temporal Data Mining, Spatio-Temporal Data Mining, Multimedia Data Mining and Phenomenal Data Mining. DM today performs computation on the database or warehouse at a single geographical location paving way for increased computation cost and questioning on data privacy. Future scope of DM is computing data located at different geographical locations. This is termed DDM/CDM (Collective Data Mining) [1].The main factors which led to the evolution of DDM are -privacy of sensitive data, transmission cost, computation cost and memory cost. The objective of DDM is to extract useful information from data located at heterogeneous sites. Distributed computing comprises distributed sites, hosting computing units at individual heterogeneous points. DDM follows decentralized mining strategy which differs from centralized strategy making entire working system scalable by distributing workload across heterogeneous sites [1].Alfredo Cuzzocrea[2] stated that framing a methodology for DDM is challenging not only by distributed environment, but also for its efficient resource sharing and minimized computational complexity specifications. DDM mainly comprises of two variations -data distributed and computation distributed. In the former method, data is distributed among heterogeneous sites at local level and computation is hosted at global level. In the latter method, computation is distributed among heterogeneous sites at local level and data is hosted at global level. Figure 1 explains DDM working architecture. The database of heterogeneous sites hosts useful, unknown information. DDM algorithms will be applied over data at heterogeneous sites as local model and finally the DM computed result will be agglomerated to form global model [1]. Kargupta et al.[3] and Zaki et al. [4] discussed that several researchers analyzed the complexity involved in framing methodology for DDM in two ways: analyzing on effective and...

show abstract

Confidence ratio affinity propagation in ensemble selection of neural network classifiers for distributed privacy-preserving data mining

Cited by 19 publications

References 42 publications

Opposition Intensity-Based Cuckoo Search Algorithm for Data Privacy Preservation

Opposition Intensity-Based Cuckoo Search Algorithm for Data Privacy Preservation

Route Map of Privacy Preservation to IOT

Approaches and Techniques of Distributed Data Mining : A Comprehensive Study

Contact Info

Product

Resources

About