Jinfei Liu

Rating: Privacy Preservation for Multiple Attributes with Different Sensitivity Requirements

Motivated by the insufficiency of the existing framework that could not process multiple attributes with different sensitivity requirements on modeling real world privacy requirements for data publishing, we present a novel method, rating, for publishing sensitive data. Rating releases AT (Attribute Table) and IDT (ID Table) based on different sensitivity coefficients for different attributes. This approach not only protects privacy for multiple sensitive attributes, but also keeps a large amount of correlations of the micro data. We develop algorithms for computing AT and IDT that obey the privacy requirements for multiple sensitive attributes, and maximize the utility of published data as well. We prove both theoretically and experimentally that our method has better performance than the conventional privacy preserving methods on protecting privacy and maximizing the utility of published data. To quantify the utility of published data, we propose a new measurement named classification measurement.

Finding Pareto optimal groups: group-based skyline

Skyline computation, aiming at identifying a set of skyline points that are not dominated by any other point, is particularly useful for multi-criteria data analysis and decision making. Traditional skyline computation, however, is inadequate to answer queries that need to analyze not only individual points but also groups of points. To address this gap, we generalize the original skyline definition to the novel group-based skyline (G-Skyline), which represents Pareto optimal groups that are not dominated by other groups. In order to compute G-Skyline groups consisting of k points efficiently, we present a novel structure that represents the points in a directed skyline graph and captures the dominance relationships among the points based on the first k skyline layers. We propose efficient algorithms to compute the first k skyline layers. We then present two heuristic algorithms to efficiently compute the G-Skyline groups: the point-wise algorithm and the unit group-wise algorithm, using various pruning strategies. The experimental results on the real NBA dataset and the synthetic datasets show that G-Skyline is interesting and useful, and our algorithms are efficient and scalable.

Privacy preserving distributed DBSCAN clustering

DBSCAN is a well-known density-based clustering algorithm which offers advantages for finding clusters of arbitrary shapes compared to partitioning and hierarchical clustering methods. However, there are few papers studying the DBSCAN algorithm under the privacy preserving distributed data mining model, in which the data is distributed between two or more parties, and the parties cooperate to obtain the clustering results without revealing the data at the individual parties. In this paper, we address the problem of two-party privacy preserving DBSCAN clustering. We first propose two protocols for privacy preserving DBSCAN clustering over horizontally and vertically partitioned data respectively and then extend them to arbitrarily partitioned data. We also provide analysis of the performance and proof of privacy of our solution.

Faster output-sensitive skyline computation algorithm

We present the second output-sensitive skyline computation algorithm which is faster than the only existing output-sensitive skyline computation algorithm [1] in worst case because our algorithm does not rely on the existence of a linear time procedure for finding medians.

Finding Probabilistic k-Skyline Sets on Uncertain Data

Skyline is a set of points that are not dominated by any other point. Given uncertain objects, probabilistic skyline has been studied which computes objects with high probability of being skyline. While useful for selecting individual objects, it is not sufficient for scenarios where we wish to compute a subset of skyline objects, i.e., a skyline set. In this paper, we generalize the notion of probabilistic skyline to probabilistic k-skyline sets (Pk-SkylineSets) which computes k-object sets with high probability of being skyline set. We present an efficient algorithm for computing probabilistic k-skyline sets. It uses two heuristic pruning strategies and a novel data structure based on the classic layered range tree to compute the skyline set probability for each instance set with a worst-case time bound. The experimental results on the real NBA dataset and the synthetic datasets show that Pk-SkylineSets is interesting and useful, and our algorithms are efficient and scalable.

Secure Skyline Queries on Cloud Platform

Outsourcing data and computation to cloud server provides a cost-e ective way to support large scale data storage and query processing. However, due to security and privacy concerns, sensitive data (e.g., medical records) need to be protected from the cloud server and other unauthorized users. One approach is to outsource encrypted data to the cloud server and have the cloud server perform query processing on the encrypted data only. It remains a challenging task to support various queries over encrypted data in a secure and e cient way such that the cloud server does not gain any knowledge about the data, query, and query result. In this paper, we study the problem of secure skyline queries over encrypted data. The skyline query is particularly important for multicriteria decision making but also presents significant challenges due to its complex computations. We propose a fully secure skyline query protocol on data encrypted using semanticallysecure encryption. As a key subroutine, we present a new secure dominance protocol, which can be also used as a building block for other queries. Finally, we provide both serial and parallelized implementations and empirically study the protocols in terms of e ciency and scalability under di erent parameter settings, verifying the feasibility of our proposed solutions.

Speed Partitioning for Indexing Moving Objects

Indexing moving objects has been extensively studied in the past decades. Moving objects, such as vehicles and mobile device users, usually exhibit some patterns on their velocities, which can be utilized for velocity-based partitioning to improve performance of the indexes. Existing velocity-based partitioning techniques rely on some kinds of heuristics rather than analytically calculate the optimal solution. In this paper, we propose a novel speed partitioning technique based on a formal analysis over speed values of the moving objects. We first formulate the optimal speed partitioning problem based on search space expansion analysis and then compute the optimal solution using dynamic programming. We then build the partitioned indexing system where queries are duplicated and processed in each index partition. Extensive experiments demonstrate that our method dramatically improves the performance of indexes for moving objects and outperforms other state-of-the-art velocity-based partitioning approaches.

Database Fragmentation with Confidentiality Constraints: A Graph Search Approach

Database fragmentation is a promising approach that can be used in combination with encryption to achieve secure data outsourcing which allows clients to securely outsource their data to remote untrusted server(s) while enabling query support using the outsourced data. Given a set of confidentiality constraints, it vertically partitions the database into fragments such that the set of attributes in each constraint do not appear together in any one fragment. The optimal fragmentation problem is to find a fragmentation with minimum cost for query support. In this paper, we propose an efficient graph search based approach which obtains near optimal fragmentation. We model the fragmentation search space as a graph and propose efficient search algorithms on the graph. We present static and dynamic search strategies as well as a novel level-wise graph expansion technique which dramatically reduces the search time. Extensive experiments showed that our method significantly outperforms other state-of-the-art methods.

Half-Plane Voronoi Diagram

In normal Voronoi diagram, each site is able to see all points in the plane. In this paper, we study the problem such that each site is only able to see half-plane and construct the so-called Half-plane Voronoi Diagram (HPVD). We show that the half-plane Voronoi cell of each site is not necessary convex and it could consist of many disjoint regions. We prove that the complexity of the HPVD of n sites is

A privacy framework: indistinguishable privacy

O(n^2)