A Textual Warehouse Approach

An enterprise memory must be able to be used as a basis for the processes of scientific or technical developments. Indeed, it was proven that information useful to these processes is not found solely in the operational bases of companies; it is also found in textual information and exchanged documents. For that reason, we propose the design and implementation of a documentary memory for business document warehouses. Its main characteristic is to allow the storage, retrieval, interrogation and analysis of information extracted from disseminated sources and, in particular, from the Web.

Multilingual Web Content Mining

This chapter presents a novel user-oriented, concept-based approach to multilingual web content mining using self-organizing maps. The multilingual linguistic knowledge required for multilingual web content mining is made available by encoding all multilingual concept-term relationships using a multilingual concept space. With this linguistic knowledge base, a concept-based multilingual text classifier is developed. It reveals the conceptual content of multilingual web documents and forms concept categories of multilingual web documents on a concept-based browsing interface. To personalize multilingual web content mining, a concept-based user profile is generated from a user’s bookmark file to highlight the user’s topics of information interest on the browsing interface. As such, both explorative browsing and user-oriented, concept-focused information filtering in multilingual web are facilitated.

A Study on Web Searching

Web search engines are one of the most popular services to help users find useful information on the Web. Although many studies have been carried out to estimate the size and overlap of the general web search engines, it may not benefit the ordinary web searching users, since they care more about the overlap of the top N (N=10, 20 or 50) search results on concrete queries, but not the overlap of the total index database. In this study, we present experimental results on the comparison of the overlap of the top N (N=10, 20 or 50) search results from AlltheWeb, Google, AltaVista and WiseNut for the 58 most popular queries, as well as for the distance of the overlapped results. These 58 queries are chosen from WordTracker service, which records the most popular queries submitted to some famous metasearch engines, such as MetaCrawler and Dogpile. We divide these 58 queries into three categories for further investigation. Through in-depth study, we observe a number of interesting results: the overlap of the top N results retrieved by different search engines is very small; the search results of the queries in different categories behave in dramatically different ways; Google, on average, has the highest overlap among these four search engines; each search engine tends to adopt a different rank algorithm independently.

Development of Agent-Based Electronic Catalog Retrieval System

Heavy electric machinery industry is currently developing electronic market places of product and parts. PLIB is the standard of dictionary model and content model for describing both commercial specifications and technical specifications of the parts and products used in the heavy electric machinery industry. This chapter represents development of an agent-based electronic catalog retrieval system using a multi-agent framework Bee-gent, in order to exchange PLIB catalog data between existing heterogenous electronic catalog servers. This chapter also gives qualitative discussion of the developed system.

Social Coordination with Architecture for Ubiquitous Agents-CONSORTS

We propose a social coordination mechanism that is realized with CONSORTS, a new kind of multi-agent architecture for ubiquitous agents. By social coordination, we mean mass users’ decision making in their daily lives, such as the mutual concession of spatial-temporal resources achieved by automatic negotiation of software agents, rather than by verbal and explicit communication directly done by human users. The prerequisite infrastructure for such an electronic negotiation mechanism is a multi-agent architecture for ubiquitous agents that are grounded in the physical world, by which software agents can trace users’ moving history, understand their intentions and preferences, and negotiate each other, all while protecting users’ privacy through temporal identifiers. The functionality of social coordination is realized in the agent architecture, where three kinds of agents work cooperatively, i.e., a personal agent that serves as proxy of the user; a social coordinator working as a service agent; and a spatio-temporal reasoner. We also summarize some basic mechanisms of social coordination functionality, including stochastic distribution and market mechanisms.

Customized Recommendation Mechanism Based on Web Data Mining and Case-Based Reasoning

One of the attractive topics in the field of Internet business is blending Artificial Intelligence (AI) techniques with the business process. In this research, we suggest a web-based, customized hybrid recommendation mechanism using Case-Based Reasoning (CBR) and web data mining. CBR mechanisms are normally used in problems for which it is difficult to define rules. In web databases, features called attributes are often selected first for mining the association knowledge between related products. Therefore, data mining is used as an efficient mechanism for predicting the relationship between goods, customers’ preference, and future behavior. If there are some goods, however, which are not retrieved by data mining, we can’t recommend additional information or a product. In this case, we can use CBR as a supplementary AI tool to recommend the similar purchase case. Web log data gathered in a real-world Internet shopping mall was given to illustrate the quality of the proposed mechanism. The results showed that the CBR and web data mining-based hybrid recommendation mechanism could reflect both association knowledge and purchase information about our former customers.

Computational Intelligence Techniques Driven Intelligent Agents for Web Data Mining and Information Retrieval

The World Wide Web has added an abundance of data and information to the complexity of information for disseminators and users alike. With this complexity has come the problem of finding useful and relevant information. There is a need for improved and intelligent search and retrieval engines. Current search engines are primarily passive tools. To improve the results returned by searches, intelligent agents and other technology have the potential, when used with existing search and retrieval engines, to provide a more comprehensive search with an improved performance. This research provides the building blocks for integrating intelligent agents with current search engines. It shows how an intelligent system can be constructed to assist in better information filtering, gathering and retrieval. The research is unique in the way the intelligent agents are directed and in how computational intelligence techniques (such as evolutionary computing and fuzzy logic) and intelligent agents are combined to improve information filtering and retrieval. Fuzzy logic is used to access the performance of the system and provide evolutionary computing with the necessary information to carry out its search.

Using Dynamically Acquired Background Knowledge for Information Extraction and Intelligent Search

This chapter presents a simple framework for extracting information found in publications or documents that are issued in large volumes and which cover similar concepts or issues within a given domain. The general aim of the work described is to present a model for automatically augmenting segments of these documents with metadata, using dynamically acquired background domain knowledge to help users easily locate information within these documents through a structured front end. To realize this goal, both document structure and dynamically acquired background knowledge are utilized. A real life example where these ideas have been applied is also presented.

An Algorithm of Pattern Match Being Fit for Mining Association Rules

There are frequent occurrences of pattern match involved in the process of counting the support count of candidates, which is one of the main factors influencing the efficiency of mining for association rules. In this chapter, an efficient algorithm for pattern match being fit for mining association rules is presented by analyzing its characters, and it has been proved correctly and efficiently.

Generating and Adjusting Web Sub-Graph Displays for Web Navigation

A graph can be used for web navigation. The whole of cyberspace can be regarded as one huge graph. To explore this huge graph, it is critical to find an effective method for tracking a sequence of the graph’s subsets (web sub-graphs) based on the user’s focus. This chapter introduces our method for generating and adjusting web sub-graph displays in the process of web navigation. Any online web sub-graph should fit in the display window. To enhance the display, there should not be any overlap between node images in the web sub-graph. Our system ensures that any online web sub-graph has no overlapping node images by letting the user, or the system itself, define the visible and invisible parts of the web graph.