It is very difficult to predict the behavior of rules set in active database. The termination of rules set can be analyzed by triggering graph (TG). It is necessary to calculate all cycles in TG to judge exactly the termination of rules set. This paper analyzes the termination of active rules set in Object - Oriented Database, produces the method of judging the termination of rules set, deduces the algorithm of calculating all cycles of triggering graph and the algorithm of judging termination of rules set.

As parts of global information infrastructure, digital libraries will likely be accessed by people all over world. Ontologies and their association can alleviate the heterogeneity and particularly the diversity of languages. This paper proposes a solution to cross-lingual information retrieval problem via ontology alignment. We elaborate two original techniques, i.e. primitives' association based on CL-LSI and mapping configuration optimization, to augment existing ontology mapping technology. As a result, multilingual collections can be bridged by this mapping, and searching across them can be achieved by three tractable steps: querying against local ontology, routing to target ontologies, and harvesting contents there online.

For many KDD applications, such as data cleaning, detecting criminal activities in E-commerce, finding the outlier can be more meaningful and interesting than finding the common cases. In the paper, we present a novel and efficient subspace local outlier test algorithm: EDOLOIS, so as to avoid the computation-intensive distance computation. The algorithm takes full use of the character of subspace data processing and the initial LOF itself, thus it can not only reduce the computation dramatically, but also gain the precise LOF of all objects in the subspaces. Both formal analysis and comprehensive performance evaluation show that the method is efficient to find all local outliers from high-dimensional categorical datasets in all subspaces.

Density estimation is a costly operation for computingdistribution information of data sets underlying many important data mining applications, such as clustering andbiased sampling. However, traditional density estimationmethods are inapplicable for streaming data, which arecontinuously arriving large volume of data, because of theirrequest for linear storage and square size calculation. Theshortcoming limits the application of many existing effective algorithms on data streams, for which the mining problem is an emergency for applications and a challenge forresearch. In this paper, the problem of computing densityfunctions over data streams is examined. A novel methodattacking this shortcoming of existing methods is developedto enable density estimation for large volume of data in linear time, fixed size memory, and without lose of accuracy .The method is based on M-Kernel merging, so that limited kernel functions to be maintained are determined intelligently. The application of the new method on differentstreaming data models is discussed, and the result of intensive experiments is presented. The analytical and empirical result show that this new density estimation algorithmfor data streams can calculate density functions on demandat any time with high accuracy for different streaming datamodels.

As a widely used data mining technique, outlier detection is a process which aims at finding anomalies with good explanations. Most existing methods are designed for numeric data. They will have problems with real-life applications that contain categorical data. In this paper, we introduce a novel outlier mining method based on a hypergraph model. Since hypergraphs precisely capture the distribution characteristics in data subspaces, this method is effective in identifying anomalies in dense subspaces and presents good interpretations for the local outlierness. By selecting the most relevant subspaces, the problem of "curse of dimensionality" in very large databases can also be ameliorated. Furthermore, the connectivity property is used to replace the distance metrics, so that the distance-based computation is not needed anymore, which enhances the robustness for handling missing-value data. The fact, that connectivity computation facilitates the aggregation operations supported by most SQL-compatible database systems, makes the mining process much efficient. Finally, experiments and analysis show that our method can find outliers in categorical data with good performance and quality.

Based on the query expansion techniques in information retrieval systems, structure-based query expansion for XML search engines, which is designed to ease the query for XML data while keeping the power and flexibility of XML query, is introduced in this paper. To enable the structure expansion, a structure thesaurus should be built first, which involves the construction of a weighted graph from XML documents and the linkage-based clustering method to cluster the nodes into several groups. After a query comes, the structure thesaurus is examined, so that for each tag in the original query, the tags in the same group are retrieved. Unrelated tags are filtered and some heuristic rules are applied to replacing the tags in the original query with the related tags and to expanding the structure. It is shown that using structure-based query expansion, the system can return result with high precision and recall.

Density estimation is an important but costly operation for applications that need to know the distribution of a data set. Moreover, when the data comes as a stream, traditional density estimation methods cannot cope with it efficiently. In this paper, we examined the problem of computing density function over data streams and developed a novel method to solve it. A new concept M-Kernel is used in our algorithm, and it is of the following characteristics: (1) the running time is in linear with the data size, (2) it can keep the whole computing in limited size of memory, (3) its accuracy is comparable to the traditional methods, (4) a useable density model could be available at any time during the processing, (5) it is flexible and can suit with different stream models. Analytical and experimental results showed the efficiency of the proposed algorithm.

Topic distillation is the process of finding representative pages relevant to a given query. Well-known topic distillation approaches such as the HITS algorithm have shown to be useful for topic distillation. Many succeeding researchers focus on augmenting HITS with further content analysis to alleviate the steady deterioration of distillation quality suffered by HITS. In this paper, we attempt to revisit the behavior of HITS from a different point of view. Namely, a similarity-based analysis model is applied to observing the distillation procedure. By defining a generalized similarity, an algorithm is proposed, which can improve the quality of distillation only using the information of hyperlinks. The experimental results reveal that the new algorithm improves distillation quality without utilizing any content information of pages.

在许多KDD(knowledge discovery in databases)应用中,如电子商务中的欺诈行为监测,例外情况或离群点的发现比常规知识的发现更有意义.现有的离群点发现大多是针对数值属性的,而且这些方法只能发现离群点,不能对其含义进行解释.提出了一种基于超图模型的离群点(outlier)定义,这一定义既体现了"局部"的概念,又能很好地解释离群点的含义.同时给出了HOT(hypergraph-based outlier test)算法,通过计算每个点的支持度、隶属度和规模偏差来检测离群点.该算法既能够处理数值属性,又能够处理类别属性.分析表明,该算法能有效地发现高维空间数据中的离群点.

In this paper, a new effective method is proposed to find class association rules (CAR), to get useful class association rules (UCAR) by removing the spurious class association rules (SCAR), and to generate exception class association rules (ECAR) for each UCAR. CAR mining, which integrates the techniques of classification and association, is of great interest recently. However, it has two drawbacks: one is that a large part of CARs are spurious and maybe misleading to users; the other is that some important ECARs are difficult to find using traditional data mining techniques. The method introduced in this paper aims to get over these flaws. According to our approach, a user can retrieve correct information from UCARs and know the influence from different conditions by checking corresponding ECARs. Experimental results demonstrate the effectiveness of our proposed approach.

The number of the measured parameters in a power plant is enormous, and the measurements are distributed all over the plant. Some must be measured at one site and used in other sites, and some must be measured and used in specific online demands. This article introduces components of distributed measurement systems, analyzes the framework of a system that was built at Southeast University, and focuses on specific power industry applications of the system. Examples include the application to a boiler accessory device strain test, the real-time query of the measured parameters of the generators and their auxiliary devices, and online video viewing of the monitored sites in a power plant's distributed environments.