Information Filtering: A New Two-Phase Model Using Stereotypic User Profiling

Computer users often experience the lost in informationspace syndrome. Information filtering suggests a solution based onrestricting the amount of information made available to users. Thisstudy suggests an advanced model for information filtering which isbased on a two-phase filtering process. The user profiling in themodel is constructed on the basis of the user's areas of interestand on sociological parameters about him that are known to thesystem. The system maintains a database of known stereotypes thatincludes rules on their information retrieval needs and habits.During the filtering process, the system relates the user to one ormore stereotypes and operates the appropriate stereotypic rules.     

基于P2P网络的信息过滤与推荐技术研究

共享信息的集中存储对存放这些信息的服务器提出了较高的要求，同时，服务器将成为整个系统的瓶颈。为此，提出了一种基于P2P的信息共享与推荐模型，解决了信息集中存放产生的问题。接着，对该模型中涉及到的基于内容的过滤，提出了一种基于词讲：链的方法，较好地解决了纯粹单一关键词无法准确描述文本的问题，并对信息推荐中使用最成功的协同过滤算法进行了描述。给出了文本过滤的实验结果及其分析。     

Alambic: a privacy-preserving recommender system for electronic commerce

Recommender systems enable merchants to assist customers in finding products that best satisfy their needs. Unfortunately, current recommender systems suffer from various privacy-protection vulnerabilities. Customers should be able to keep private their personal information, including their buying preferences, and they should not be tracked against their will. The commercial interests of merchants should also be protected by allowing them to make accurate recommendations without revealing legitimately compiled valuable information to third parties. We introduce a theoretical approach for a system called Alambic, which achieves the above privacy-protection objectives in a hybrid recommender system that combines content-based, demographic and collaborative filtering techniques. Our system splits customer data between the merchant and a semi-trusted third party, so that neither can derive sensitive information from their share alone. Therefore, the system could only be subverted by a coalition between these two parties.

Slicing, Chopping, and Path Conditions with Barriers

One of the critiques on program slicing is that slices presented to the user are hard to understand. This is mainly related to the problem that slicing dumps the results onto the user without any explanation. This work will present an approach that can be used to filter slices. This approach basically introduces barriers which are not allowed to be passed during slice computation. An earlier filtering approach is chopping which is also extended to obey such a barrier. The barrier variants of slicing and chopping provide filtering possibilities for smaller slices and better comprehensibility. The concept of barriers is then applied to path conditions, which provide necessary conditions under which an influence between the source and target criterion exists. Barriers make those conditions more precise.     

基于搜索结果的个性化推荐系统研究

搜索引擎返回的信息太多且不能根据用户的兴趣提供检索结果,使得用户使用搜索引擎难以用简便的方式找到感兴趣的文档。个性化推荐是一种旨在减轻用户在信息检索方面负担的有效方法。文中把内容过滤技术和文档聚类技术相结合,实现了一个基于搜索结果的个性化推荐系统,以聚类的方法自动组织搜索结果,主动推荐用户感兴趣的文档。通过建立用户概率兴趣模型,对搜索结果STC聚类的基础上进行内容过滤。实验表明,概率模型比矢量空间模型更好地表达了用户的兴趣和变化。     

A best practice advice system to support automotive engineering analysis processes

Engineering design teams today are often widely distributed, and design authority is shared between collaborating companies. Technology is changing rapidly, and understanding of the most appropriate approach to the application of engineering assessment tools is developing accordingly. There is therefore a need to support coordination and auditing of engineering processes, and to provide best practice advice. This paper describes a computing approach to the provision of best practice advice within a workflow-enabled engineering computing environment. The engineering context is described using a formal information model for automotive engineering analysis processes, embedded in an object database. This same model is used to associate best practice advice documents with the engineering context. The best practice adviser (BPA) system assembles four types of information: general information that is pertinent to a particular activity, irrespective of the context in which it is taking place; context-specific information that is pertinent to the particular circumstance in which an activity is taking place; errors and warnings that may be encountered in the activity, especially when software is being used, and examples of previous application of the activity in related contexts. The BPA is implemented in a three-tier architecture using server pages technology. In the absence of any suitable matching information for a particular context in the BPA database, the BPA Server can execute a close-match algorithm which searches the database for information that is provided on contexts that are close to the users interest. The paper describes the initial implementation and population of the BPA, and presents some early feedback from prototype trials.     

Authoring SMIL documents by direct manipulations during presentation

This paper presents SmilEditor – an authoring environment to write multimedia documents encoded in SMIL. The main feature of SmilEditor is to strongly integrate the presentation view, in which the document is executed, with the editing process. In this view objects can be selected to perform a wide set of editing actions ranging from attributes setting to direct spatial or temporal editions. This way of editing a multimedia document is close to the wellknown WYSIWYG paradigm used by usual wordprocessors. Moreover, in order to help the author to specify the temporal organisation of documents, SmilEditor provides an execution report displayed through a timeline view. This view also contains information which helps the author to understand why such execution occurred. These multiple and synchronised views aim at covering the various needs for authoring multimedia documents.     

A new approach for combining content-based and collaborative filters

With the development of e-commerce and the proliferation of easily accessible information, recommender systems have become a popular technique to prune large information spaces so that users are directed toward those items that best meet their needs and preferences. A variety of techniques have been proposed for performing recommendations, including content-based and collaborative techniques. Content-based filtering selects information based on semantic content, whereas collaborative filtering combines the opinions of other users to make a prediction for a target user. In this paper, we describe a new filtering approach that combines the content-based filter and collaborative filter to capitalize on their respective strengths, and thereby achieves a good performance. We present a series of recommendations on the selection of the appropriate factors and also look into different techniques for calculating user-user similarities based on the integrated information extracted from user profiles and user ratings. Finally, we experimentally evaluate our approach and compare it with classic filters, the result of which demonstrate the effectiveness of our approach.     

A Model of Collaborating Agents for Content-Based Electronic Document Filtering

The huge volume of distributed information that is nowadays available in electronic multimedia documents forces a lot of people to consume a significant percentage of their time looking for documents that contain information useful to them. The filtering of electronic documents seems hard to automate, partly because of document heterogeneity, but mainly because it is difficult to train computers to have an understanding of the contents of these documents and make decisions based on user-subjective criteria. In this paper, we suggest a model for the automation of content-based electronic document filtering, supporting multimedia documents in a wide variety of forms. The model is based on multi-agent technology and utilizes an adaptive knowledge base organized as a set of logical rules. Implementations of the model using the client-server architecture should be able to efficiently access documents distributed over an intranet or the Internet.     

A Multilingual Text Mining Approach Based on Self-Organizing Maps

This paper describes our work on developing a language-independent technique for discovery of implicit knowledge from multilingual information sources. Text mining has been gaining popularity in the knowledge discovery field, particularity with the increasing availability of digital documents in various languages from all around the world. However, currently most text mining tools mainly focus only on processing monolingual documents (particularly English documents): little attention has been paid to apply the techniques to handle the documents in Asian languages, and further extend the mining algorithms to support the aspects of multilingual information sources. In this work, we attempt to develop a language-neutral method to tackle the linguistics difficulties in the text mining process. Using a variation of automatic clustering techniques, which apply a neural net approach, namely the Self-Organizing Maps (SOM), we have conducted several experiments to uncover associated documents based on a Chinese corpus, Chinese-English bilingual parallel corpora, and a hybrid Chinese-English corpus. The experiments show some interesting results and a couple of potential paths for future work in the field of multilingual information discovery. Besides, this work is expected to act as a starting point for exploring the impacts on linguistics issues with the machine-learning approach to mining sensible linguistics elements from multilingual text collections.     

THESUS: Organizing Web document collections based on link semantics

The requirements for effective search and management of the WWW are stronger than ever. Currently Web documents are classified based on their content not taking into account the fact that these documents are connected to each other by links. We claim that a pages classification is enriched by the detection of its incoming links semantics. This would enable effective browsing and enhance the validity of search results in the WWW context. Another aspect that is underaddressed and strictly related to the tasks of browsing and searching is the similarity of documents at the semantic level. The above observations lead us to the adoption of a hierarchy of concepts (ontology) and a thesaurus to exploit links and provide a better characterization of Web documents. The enhancement of document characterization makes operations such as clustering and labeling very interesting. To this end, we devised a system called THESUS. The system deals with an initial sets of Web documents, extracts keywords from all pages incoming links, and converts them to semantics by mapping them to a domains ontology. Then a clustering algorithm is applied to discover groups of Web documents. The effectiveness of the clustering process is based on the use of a novel similarity measure between documents characterized by sets of terms. Web documents are organized into thematic subsets based on their semantics. The subsets are then labeled, thereby enabling easier management (browsing, searching, querying) of the Web. In this article, we detail the process of this system and give an experimental analysis of its results.Received: 16 December 2002, Accepted: 16 April 2003, Published online: 17 September 2003     

Anatomy and Empirical Evaluation of an Adaptive Web-Based Information Filtering System

A case study in adaptive information filtering systems for the Web is presented. The described system comprises two main modules, named HUMOS and WIFS. HUMOS is a user modeling system based on stereotypes. It builds and maintains long term models of individual Internet users, representing their information needs. The user model is structured as a frame containing informative words, enhanced with semantic networks. The proposed machine learning approach for the user modeling process is based on the use of an artificial neural network for stereotype assignments. WIFS is a content-based information filtering module, capable of selecting html/text documents on computer science collected from the Web according to the interests of the user. It has been created for the very purpose of the structure of the user model utilized by HUMOS. Currently, this system acts as an adaptive interface to the Web search engine ALTA VISTA^TM. An empirical evaluation of the system has been made in experimental settings. The experiments focused on the evaluation, by means of a non-parametric statistics approach, of the added value in terms of system performance given by the user modeling component; it also focused on the evaluation of the usability and user acceptance of the system. The results of the experiments are satisfactory and support the choice of a user model-based approach to information filtering on the Web.     

Construction of a classifier using AdaBoost for information filtering

We describe an information filtering system using AdaBoost. To realize the filtering system, we created a user profile which presents the users interests. Since the users interests are complex, the user profile becomes a nonlinear discriminant function. However, it is difficult to decide on an appropriate discriminant function. We used AdaBoost to modify the appropriate user profile. AdaBoost is an ensemble algorithm which combines weak learners and improves the accuracy of a classifier. In this method, the weak learners for AdaBoost is a linear discriminant function which is created with a genetic algorithm. We carried out experiments for an information filtering service on an NTCIR2 test collection, and we discuss the effectiveness of the method.This work was presented, in part, at the 9th International Symposium on Artificial Life and Robotics, Oita, Japan, January 28–30, 2004     

Graph vs. bag representation models for the topic classification of web documents

Text classification constitutes a popular task in Web research with various applications that range from spam filtering to sentiment analysis. In this paper, we argue that its performance depends on the quality of Web documents, which varies significantly. For example, the curated content of news articles involves different challenges than the user-generated content of blog posts and Social Media messages. We experimentally verify our claim, quantifying the main factors that affect the performance of text classification. We also argue that the established bag-of-words representation models are inadequate for handling all document types, as they merely extract frequent, yet distinguishing terms from the textual content of the training set. Thus, they suffer from low robustness in the context of noisy or unseen content, unless they are enriched with contextual, application-specific information. In their place, we propose the use of n-gram graphs, a model that goes beyond the bag-of-words representation, transforming every document into a graph: its nodes correspond to character or word n-grams and the co-occurring ones are connected by weighted edges. Individual document graphs can be combined into class graphs and graph similarities are employed to position and classify documents into the vector space. This approach offers two advantages with respect to bag models: first, classification accuracy increases due to the contextual information that is encapsulated in the edges of the n-gram graphs. Second, it reduces the search space to a limited set of robust, endogenous features that depend on the number of classes, rather than the size of the vocabulary. Our thorough experimental study over three large, real-world corpora confirms the superior performance of n-gram graphs across the main types of Web documents.     

S-Conart: an interaction method that facilitates concept articulation in shopping online

This study addresses building an interactive system that effectively prompts customers to make their decision while shopping online. It is especially targeted at purchasing as concept articulation where customers initially have a vague concept of what they want and then gradually clarify it in the course of interaction, which has not been covered by traditional online shopping systems. This paper proposes information presentation methods to effectively facilitate customers in their concept articulation process, and the framework for interaction design to enable the methods. Specifically, this study builds a system called S-Conart that facilitates purchasing as concept articulation through support for customers conception with spatial-arrangement style information presentation and for their conviction with scene information presentation, and then makes a set of evaluation experiments with the system to verify that the approach used in building the system is effective in facilitating the purchasing as concept articulation.     

Personalized Course Navigation Based on Grey Relational Analysis

Web personalization has quickly changed from a value-added facility to a service required in presenting large quantities of information because individual users of the Internet have various needs and preferences in seeking information. This paper presents a novel personalized recommendation system with online preference analysis in a distance learning environment called Coursebot. Users can both browse and search for course materials by using the interface of Coursebot. Moreover, the proposed system includes appropriate course materials ranked according to a users interests. In this work, an analysis measure is proposed to combine typical grey relational analysis and implicit rating, and thus a users interests are calculated from the content of documents and the users browsing behavior. This algorithms low computational complexity and ease of adding knowledge support online personalized analysis. In addition, the user profiles are dynamically revised to provide efficiently personalized information that reflects a users interests after each page is visited.     

Specifying the apple Macintosh™ Toolbox event manager

The Macintosh Toolbox is a collection of resources, stored in ROM, of use to software developers building user interface components. This paper documents an attempt to specify one part of the Toolbox: namely the data types and routines for event-handling. The chief objective of the paper is to illustrate an approach to building a formal specification in which, starting from simple structures, the final specification is arrived at by a process of successive modifications, generalisations, instantiations and translations. The paper also constitutes an example of the Larch algebraic style, as well as the two-tiered approach characteristic of the Larch family of specification languages.This work has been undertaken as part of the Alvey HI 059 project Formal Aspects of Interactive Dialogues, supported under SERC Grant No. GR/D/1646.8.     

A hybrid recommender system using artificial neural networks

In the context of recommendation systems, metadata information from reviews written for businesses has rarely been considered in traditional systems developed using content-based and collaborative filtering approaches. Collaborative filtering and content-based filtering are popular memory-based methods for recommending new products to the users but suffer from some limitations and fail to provide effective recommendations in many situations. In this paper, we present a deep learning neural network framework that utilizes reviews in addition to content-based features to generate model based predictions for the business-user combinations. We show that a set of content and collaborative features allows for the development of a neural network model with the goal of minimizing logloss and rating misclassification error using stochastic gradient descent optimization algorithm. We empirically show that the hybrid approach is a very promising solution when compared to standalone memory-based collaborative filtering method.     

An evaluation of semantic fisheye views for opportunistic search in an annotated image collection

Visual interfaces are potentially powerful tools for users to explore a representation of a collection and opportunistically discover information that will guide them toward relevant documents. Semantic fisheye views (SFEVs) are focus + context visualization techniques that manage visual complexity by selectively emphasizing and increasing the detail of information related to the users focus and deemphasizing or filtering less important information.In this paper we describe a prototype for visualizing an annotated image collection and an experiment to compare the effectiveness of two distinctly different SFEVs for a complex opportunistic search task. The first SFEV calculates relevance based on keyword-content similarity and the second based on conceptual relationships between images derived using WordNet. The results of the experiment suggest that semantic-guided search is significantly more effective than similarity-guided search for discovering and using domain knowledge in a collection.