Advances in Scalable Web Information Integration and Service

The following sections are included:

• Workshop Organizers

• Preface

• Table of Contents

XML is emerging as a de facto standard for information exchange over the Web, while businesses and enterprises generate and exchange large amounts of XML data daily. One of the major challenges is how to query this data efficiently. Queries typically can be represented as twig patterns. Some researchers have developed algorithms that reduce the intermediate results that are generated during query processing, while others have introduced labeling schemes that encode the position of elements, enabling queries to be answered by accessing the labels without traversing the original XML documents. In this paper we outline optimizations that are based on semantics of the data being queried, and introduce efficient algorithms for content and keyword searches in XML databases. If the semantics are known we can further optimize the query processing, but if the semantics are unknown we revert to the traditional query processing approaches.

We propose a method for processing aggregate queries by analyzing indices of Web search engines. An aggregate query is a form of query such as "What are typical topics on Kyoto?" and the answer is a set of pairs of a topic term and its statistical score. Using search engine indices to process aggregate queries has two advantages. One is that we can realize quick and instant processing by using indices and minimizing the access to actual contents. The other is that results are statistically reliable since they are based on a huge amount of Web data. Our method consists of two phases. The probe phase returns a set of terms that satisfy a certain relation to a given keyword. The validation phase determines whether given terms satisfy a certain relation. We describe examples of aggregate queries: finding typical topics, finding siblings/rivals, and finding landmarks.

Web service composition provides a way to build value-added services and web applications by integrating and composing existing web services. A composite web service is essential a process in a loosely-coupled service-oriented architecture. To conduct performance analysis, a workflow representation of the underlying process is required. This paper describes a method to discover such underlying processes from history execution logs. The workflow logs contain temporal information that records the starting and ending time of activities. Based on a probabilistic model, the algorithm can discover sequential, parallel, exclusive choice and iterative structures. Some examples are given to illustrate the algorithm. Though discussed process mining of a composite web service, this method can also be used for mining other workflow applications.

In Semantic Web search, the resource type that does not exist in the ontology becomes to be a big issue. In this paper, we propose a Web resource categorization method for Semantic Web Search to search the classes that have high relation to the user's intent. Our proposed categorization method infers relations between classes even without explicit definitions, by using the construction of classes in ontologies and the associations between classes and properties in RDFs. Finally, we confirm the effectivity of the proposed approach by comparing with conventional approaches.

Query on deep web has been gaining great importance due to the large amount of deep web data sources. Inspired by the observations on deep web, the paper presents a novel top-k ranking strategy to rank the relevant data sources according to the user's requirement. First, it applies an attribute based frequent pattern growth algorithm to mining the most dominant attributes, and then employs a dominance and relevance based top-k style ranking algorithm to find the relevant data sources with early candidates pruning and termination, which has proved high efficiency. We have conducted extensive experiments on a real world dataset and demonstrated the efficiency of our approach.

As Web Services become increasingly important so do the portals that index the services. Common portal of Web Services require publishing to the registry. While manually compiled registries are useful, they often fail to include a large number of Web Services available for public use. Rather than require a Web Service operator to publish to a registry or catalog, we propose to create an automated discovery system which finds and indexes Web Services within a particular application. By using a combination of directed search and web crawling we are able to find a large number of candidate sites which can be validated and added to a portal. Results show that our method is successful in discovering large number of Geospatial Web Services and is better than the methods which rely solely on directed search.

With the rapid increase of Web Service sites, one emerging problem is how to find most appropriate Services among massive number of similar Web Services and another problem is how to overcome the heterogeneity of metadata among them. In this paper, we propose intelligent service finding techniques for the HERMES, which is a Web Service-enabled visual media retrieval framework. We show how Semantic Web technology can be utilized to find appropriate visual media Web Services. We also propose how metadata can be transformed to overcome the heterogeneity among visual media Web Services.

In recent years, the Web has become a popular medium for disseminating information, news, ideas, and opinions of the modern society. Due to this phenomenon, the Web information is reflecting current events and trends that are happening in the real world which, in turn, has attracted a lot of interest in using the Web as a sociological research tool for detecting the emerging topics, and social trends. To facilitate such kind of sociological research, in this paper, we study the characteristics of socio-topical web keywords sampled from a series of Thai web snapshots. The socio-topical web keyword, extracted from the content of some web pages, is a keyword relating to some topics of interest in a real-world society. The study was conducted as follows. First, the socio-topical keywords were sampled from the inverted index of each Thai web snapshot. Then, for each sampled keyword, we observe the pattern of changes of the number of documents containing the keyword, and the inverse document frequency (IDF) scores. Finally, we try to find the relationships between the observed patterns of changes and their corresponding real-world events in the Thai society.

Integrating various data sources is a major problem in knowledge management. Integration systems proposed in the last two decades as a solution start showing some limitations for various reasons. First the environment scaled from a few data sources in the same location to an unknown number of possible data sources on the Web. Second the sources to be integrated may change quickly over time. We propose in this article a method allowing for the dynamic integration of data sources on the web. Based on a network of schemas (database schemas, ontologies) related via mappings our algorithm generates a global view over a set of resources. Our approach presents the advantage to minimize the necessary human intervention in order to integrate sources.

A challenging issue in web is the integration of XML data sources. In this paper we address this issue when the XML data sources conform to schemas. This is a challenging problem because the XML documents may have significant structural differences even if they represent data from the same knowledge domain. We deal with it by (a) adding semantic information to the document schemas that allows the grouping of schema elements into the so called schema dimensions, and (b)using a query language that allows a partial specification of tree patterns. We show how the schema dimensions can form a dimension graph for all the data sources that acts as a global schema. Users posing their queries on this global schema have the flexibility of specifing structural constraints fully, partially or not at all. Our approach allows querying data sources with different schemas in an integrated way. We show how these queries can be translated into sets of branching paths expressions to be evaluated on the XML documents of the data sources. Our approach can be easily implemented on top of an XQuery engine.

XML has become a standard for data representation and exchange on the Internet. The functional dependency is the most important part of the research on XML database normalization and there is already some research work on this topic. According to the types of existing complicated functional dependency constraint, the features of the data constraint are analyzed. A definition of XML functional dependency (XDF) is proposed which is based on a DTD path language. Compared with formal definitions already proposed, XFD proposed in this paper can express more XML dependencies, which makes it more suitable for further research work.

The following sections are included:

• Author Index