一种改进的结合K近邻法的SVM分类算法

在对支持向量机在超平面附近容易对测试样本造成错分进行研究的基础上,改进了将支持向量机分类和k近邻分类相结合的方法,形成了一种新的分类器.在分类阶段计算待识别样本和最优分类超平面的距离,如果距离差大于给定阈值可直接应用支持向量机分类,否则用最佳距离k近邻分类.数值实验表明,使用支持向量机结合最近邻分类的分类器分类比单独使用支持向量机分类具有更高的分类准确率.     

基于马氏距离和灰色分析的缺失值填充算法

针对kNN算法中欧氏距离具有密度相关性敏感的缺点,提出综合马氏距离和灰色分析方法代替kNN算法中欧式距离的新算法,应用到缺失数据填充方面。其中马氏距离能解决密度相关明显的数据集,灰色分析方法能处理密度相关不明显的情况。因此,该算法能很好处理任何数据集,实验结果显示,算法在填充结果上明显优于现有的其他算法。     

Protecting location privacy in mobile geoservices using fuzzy inference systems

Mobile geoservices, especially location-based services (LBSs), are becoming more popular each day. The most important goal of these services is to use a user’s location to provide location-aware services. Because the user’s spatial information can be abused by organizations or advertisers, and sometimes for criminal purposes, the protection of this information is a necessary part of such services. There has been substantial research on privacy protection in LBSs and mobile geoservices; most studies have attempted to anonymize the user and hide his/her identity or to engage the user in the protection process. The major defects of these previous approaches include an increased complexity of system architecture, a decrease in service capabilities, undesirable processing times, and a failure to satisfy users. Additionally, anonymization is not a suitable solution for context-aware services. Therefore, in this paper, a new approach is proposed to locate users with different levels of spatial precision, based on his/her spatio-temporal context and a user’s group, through fuzzy inference systems. The user’s location and the time of the request determine the spatio-temporal context of the user. A fuzzy rule base is formed separately for each group of users and services. An interview is a simple method to extract the rules. The spatial precision of a user’s location, which is obtained from a fuzzy system, goes to a spatial function called the conceptualization function, to determine the user’s location based on one of the following five levels of qualitative precision: geometrical coordinates, streets, parish, region, and qualitative location, such as the eastern part of the city. Thus, there is no need to anonymize users in mobile geoservices or to turn the service off. The applicability and efficiency of the proposed method are shown for a group of taxi drivers.     

A crowd enabled approach for processing nearest neighbor and range queries in incomplete databases with accuracy guarantee

With the proliferation of mobile devices and wireless technologies, location based services (LBSs) are becoming popular in smart cities. Two important classes of LBSs are Nearest Neighbor (NN) queries and range queries that provide user information about the locations of point of interests (POIs) such as hospitals or restaurants. Answers of these queries are more reliable and satisfiable if they come from trustworthy crowd instead of traditional location service providers (LSPs). We introduce an approach to evaluate NN and range queries with crowdsourced data and computation that eliminates the role of an LSP. In our crowdsourced approach, a user evaluates LBSs in a group. It may happen that group members do not have knowledge of all POIs in a certain area. We present efficient algorithms to evaluate queries with accuracy guarantee in incomplete databases. Experiments show that our approach is scalable and incurs less computational overhead.     

On multi-type reverse nearest neighbor search

This paper presents a study of the Multi-Type Reverse Nearest Neighbor (MTRNN) query problem. Traditionally, a reverse nearest neighbor (RNN) query finds all the objects that have the query point as their nearest neighbor. In contrast, an MTRNN query finds all the objects that have the query point in their multi-type nearest neighbors. Existing RNN queries find an influence set by considering only one feature type. However, the influence from multiple feature types is often critical for strategic decision making in many business scenarios, such as site selection for a new shopping center. To that end, we first formalize the notion of the MTRNN query by considering the influence of multiple feature types. We also propose R-tree based algorithms to find the influence set for a given query point and multiple feature types. Finally, experimental results are provided to show the strength of the proposed algorithms as well as design decisions related to performance tuning.     

Search on transportation networks for location-based service

The issue of how to provide location-based service (LBS) attracted many researchers. In this paper, we focus on a typical situation of LBS which is to provide services for users in cars that move in a road network. To provide such kind of services, an integration method for representing transportation information with a road map is proposed. By using our integration method, since the transportation information of road networks is managed under the spatial index structure created for road networks, spatial queries on them can take advantages of the spatial index structure and achieve an efficient process. Moreover, we discuss path search, region search, nearest neighbor search and continuous nearest neighbor search in this paper, which are based on transportation networks with (or without) considering the static spatial objects outside the transportation networks. Using transportation information and the corresponding real road network, the paper offers evaluations by comparing our representation method and query method with those in related works. The results show a good performance of our methods.     

Taking advantage of collective knowledge in emergency response systems

When large groups work on a theme, they have the potential to produce a lot of useful knowledge, regardless of whether they are acting in a coordinated manner or individually. Spontaneously generated information has received much attention in recent years, as organizations and businesses discover the power of crowds. New technologies, such as blogs, Twitter, wikis, photo sharing, collaborative tagging and social networking sites, enable the creation and dissemination of content in a relatively simple way. As a result, the aggregate body of knowledge is growing at an accelerated rate. Many organizations are looking for ways to harness this power, which is being called collective intelligence. Research has shown that it is possible to obtain high quality results from collectively produced work.In this paper, we consider the domain of emergency response. Research has shown that individuals respond quickly and massively to emergencies, and that they try to help with the situation. Thus, it seems like a logical step to attempt to harness collective knowledge for emergency management. Disaster relief groups and field command frequently suffer from lack of up to date information, which may be critical in a rapidly evolving situation. Some of this information could be generated by the crowd at large, enabling more effective response to the situation. In this paper, we discuss the possibilities for the introduction of collective knowledge in disaster relief and present architecture and examples of how this could be accomplished.     

A pure peer-to-peer approach for kNN query processing in mobile ad hoc networks

Together with advanced positioning and mobile technologies, P2P query processing has attracted a growing interest number of location-aware applications such as answering kNN queries in mobile ad hoc networks. It not only overcomes drawbacks of centralized systems, for example single point of failure and bottleneck issues, but more importantly harnesses power of peers’ collaboration. In this research, we propose a pure mobile P2P query processing scheme which primarily focuses on the search and validation algorithm for kNN queries. The proposed scheme is designed for pure mobile P2P environments with the absence of the base station support. Compared with centralized and hybrid systems, our system can reduce energy consumption more than six times by making use of data sharing from peers in a reasonable mean latency of processing time for networks with high density of moving objects as can be seen in the simulation results.     

Rewriting systems with limited distance forbidding context

A random context production has a permitting and forbidding context. A symbol can be rewritten using such a production if all the permitting context symbols and no forbidding context symbol appear in the sentential string. In this paper we limit the effect of forbidding context symbols to be within a certain distance from the symbol to be rewritten. Outside this distance the forbidding context symbols do not influence the rewriting of a symbol. This restriction strictly increases the generating power of the rewriting system.

A further result of this paper is a “negative parallel” version of Penttonen's normal form.     

Rewriting systems with limited distance permitting context

By restricting the permitting context symbols in a rewriting system to be within a specified distance from the symbol to be replaced, we strictly increase the generative power above that of rewriting systems where the context symbols can appear within arbitrary distances from the symbol to be replaced.     

Privacy-preserving kNN query processing algorithms via secure two-party computation over encrypted database in cloud computing

The Journal of Supercomputing - Since studies on privacy-preserving database outsourcing have been spotlighted in a cloud computing, databases need to be encrypted before being outsourced to the...     

Dimensionality reduction for similarity search with the Euclidean distance in high-dimensional applications

In multimedia information retrieval, multimedia data are represented as vectors in high-dimensional space. To search these vectors efficiently, a variety of indexing methods have been proposed. However, the performance of these indexing methods degrades dramatically with increasing dimensionality, which is known as the dimensionality curse. To resolve the dimensionality curse, dimensionality reduction methods have been proposed. They map feature vectors in high-dimensional space into vectors in low-dimensional space before the data are indexed. This paper proposes a novel method for dimensionality reduction based on a function that approximates the Euclidean distance based on the norm and angle components of a vector. First, we identify the causes of, and discuss basic solutions to, errors in angle approximation during the approximation of the Euclidean distance. Then, this paper propose a new method for dimensionality reduction that extracts a set of subvectors from a feature vector and maintains only the norm and the approximated angle for every subvector. The selection of a good reference vector is crucial for accurate approximation of the angle component. We present criteria for being a good reference vector, and propose a method that chooses a good reference vector. Also, we define a novel distance function using the norm and angle components, and formally prove that the distance function consistently lower-bounds the Euclidean distance. This implies information retrieval with this function does not incur any false dismissals. Finally, the superiority of the proposed approach is verified via extensive experiments with synthetic and real-life data sets.

Modeling and mining of dynamic trust in complex service-oriented systems

The global scale and distribution of companies have changed the economy and dynamics of businesses. Web-based collaborations and cross-organizational processes typically require dynamic and context-based interactions between people and services. However, finding the right partner to work on joint tasks or to solve emerging problems in such scenarios is challenging due to scale and temporary nature of collaborations. Furthermore, actor competencies evolve over time, thus requiring dynamic approaches for their management. Web services and SOA are the ideal technical framework to automate interactions spanning people and services. To support such complex interaction scenarios, we discuss mixed service-oriented systems that are composed of both humans and software services, interacting to perform certain activities. As an example, consider a professional online support community consisting of interactions between human participants and software-based services. We argue that trust between members is essential for successful collaborations. Unlike a security perspective, we focus on the notion of social trust in collaborative networks. We show an interpretative rule-based approach to enable humans and services to establish trust based on interactions and experiences, considering their context and subjective perceptions.     

kNN在文本分类中的应用研究

随着网络技术与数字图书馆的迅猛发展,在线文档迅速增加,自动文本分类已成为处理和组织大量文档数据的关键技术。kNN方法作为一种简单、有效、非参数的分类方法,在文本分类中得到广泛的应用。本文介绍了kNN分类算法的思想以及两种不同的决策规则,并通过实现的文本分类系统对基于离散值规则的kNN方法和基于相似度加权的kNN方法进行实验比较。实验结果表明。基于相似度加权的kNN方法的分类性能要优于基于离散值规则的kNN方法。     

Efficient processing of graph similarity queries with edit distance constraints

Graphs are widely used to model complicated data semantics in many applications in bioinformatics, chemistry, social networks, pattern recognition, etc. A recent trend is to tolerate noise arising from various sources such as erroneous data entries and find similarity matches. In this paper, we study graph similarity queries with edit distance constraints. Inspired by the $q$ -gram idea for string similarity problems, our solution extracts paths from graphs as features for indexing. We establish a lower bound of common features to generate candidates. Efficient algorithms are proposed to handle three types of graph similarity queries by exploiting both matching and mismatching features as well as degree information to improve the filtering and verification on candidates. We demonstrate the proposed algorithms significantly outperform existing approaches with extensive experiments on real and synthetic datasets.     

Tree edit distance: Robust and memory-efficient

Hierarchical data are often modelled as trees. An interesting query identifies pairs of similar trees. The standard approach to tree similarity is the tree edit distance, which has successfully been applied in a wide range of applications. In terms of runtime, the state-of-the-art algorithm for the tree edit distance is RTED, which is guaranteed to be fast independent of the tree shape. Unfortunately, this algorithm requires up to twice the memory of its competitors. The memory is quadratic in the tree size and is a bottleneck for the tree edit distance computation.In this paper we present a new, memory efficient algorithm for the tree edit distance, AP-TED (All Path Tree Edit Distance). Our algorithm runs at least as fast as RTED without trading in memory efficiency. This is achieved by releasing memory early during the first step of the algorithm, which computes a decomposition strategy for the actual distance computation. We show the correctness of our approach and prove an upper bound for the memory usage. The strategy computed by AP-TED is optimal in the class of all-path strategies, which subsumes the class of LRH strategies used in RTED. We further present the AP-TED⁺ algorithm, which requires less computational effort for very small subtrees and improves the runtime of the distance computation. Our experimental evaluation confirms the low memory requirements and the runtime efficiency of our approach.     

Integration of knowledge-based systems with data processing

Transaction Oriented Prolog-Online Environment is a multi-user, mainframe implementation of the logic programming language Prolog, specially designed and developed for the data processing environment. This article describes the background, functionality, implementation and use of TOP-ONE, stressing that applications of knowledge-based systems technology to data processing environments will achieve full success only when existing systems and existing conventional data are incorporated within the applications.