Catálogo de publicaciones - libros

Compartir en
redes sociales

Privacy Preserving Data Mining

Jaideep Vaidya Yu Michael Zhu Christopher W. Clifton

Resumen/Descripción – provisto por la editorial

No disponible.

Palabras clave – provistas por la editorial

No disponibles.

Disponibilidad

Institución detectada	Año de publicación	Navegá	Descargá	Solicitá
No detectada	2006	SpringerLink

Información

Tipo de recurso:

libros

ISBN impreso

978-0-387-25886-7

ISBN electrónico

978-0-387-29489-6

Editor responsable

Springer Nature

País de edición

Reino Unido

Fecha de publicación

2006

Información sobre derechos de publicación

Cobertura temática

Ciencias de la computación e información

Tabla de contenidos

Verificá que desde tu institución tengas acceso para descargar o solicitar el libro completo o alguno de sus capítulos.

doi: 10.1007/978-0-387-29489-6_1

Privacy and Data Mining

Palabras clave: Data Mining; Association Rule; Private Information; Data Warehouse; Privacy Risk.

Pp. 1-5

doi: 10.1007/978-0-387-29489-6_2

What is Privacy?

Palabras clave: Data Mining; Sensitive Data; Conditional Entropy; Privacy Risk; Differential Entropy.

Pp. 7-15

doi: 10.1007/978-0-387-29489-6_3

Solution Approaches / Problems

Pp. 17-27

doi: 10.1007/978-0-387-29489-6_4

Predictive Modeling for Classification

We have seen several examples of how to create privacy-preserving classification algorithms. A common theme is to start with an algorithm and a partitioning of data. An appropriate model for preserving privacy is then chosen; If the data comes from many sources and noise is deemed sufficient to protect privacy, the data is perturbed. With fewer sources, cryptographic approaches are more appropriate. In either case, the algorithm is analyzed to determine what steps can be computed easily, and which are affected by privacy. With data perturbation, the problem is generally determining distributions; the effect of noise on individual values will generally average out over time. With cryptographic protocols, the problem is generally computing a function from distributed inputs; through clever transformations to that function and use of a limited toolkit of techniques, it is often possible to efficiently evaluate or approximate the function without disclosing data. The number of algorithms that have been developed for privacy-preserving classification continues to grow. The methods presented here capture some of the key concepts that are being used in these algorithms.

Palabras clave: Information Gain; Numeric Attribute; Class Site; Nominal Attribute; Fraud Detection.

Pp. 29-52

doi: 10.1007/978-0-387-29489-6_5

Predictive Modeling for Regression

Palabras clave: Secure Protocol; Residual Plot; Cross Term; Data Owner; Standard Statistical Metropolitan Area.

Pp. 53-69

doi: 10.1007/978-0-387-29489-6_6

Finding Patterns and Rules (Association Rules)

Palabras clave: Association Rule; Frequent Itemsets; Association Rule Mining; Find Pattern; Support Count.

Pp. 71-83

doi: 10.1007/978-0-387-29489-6_7

Descriptive Modeling (Clustering, Outlier Detection)

Pp. 85-111

doi: 10.1007/978-0-387-29489-6_8

Future Research - Problems remaining

Palabras clave: Data Mining; Credit Card; Data Mining Process; Credit Card Number; Card Transaction.

Pp. 113-114