We explore the notion of a tour guide software agent for assisting users browsing the World Wide Web. A Web tour guide agent provides assistance similar to that provided by ahuman tour guide in a museum -- it guides the user along an appropriate path through the collection, based on its knowledge of the user's interests, of the location and relevance of various items in the collection, and of the way in which others have interacted with the collection in the past. This paper describes a simple but operational tour guide, called Web-Watcher, which has given over 5000 tours to people browsing CMU's School of Computer Science Web pages. WebWatcher accompanies users from page to page, suggests appropriate hyperlinks, and learns from experience to improve its advice-giving skills. We describe the learning algorithms used by WebWatcher, experimental results showing their effectiveness, and lessons learned from this case study in Web tour guide agents.

Most web pages are linked to others with related content. This idea, combined with another that says that text in, and possibly around, HTML anchors describe the pages to which they point, is the foundation for a usable WorldWide Web. In this paper, we examine to what extent these ideas hold by empirically testing whether topical locality mirrors spatial locality of pages on the Web. In particular, we find that the likelihood of linked pages having similar textual content to be high; the similarity of sibling pages increases when the links from the parent are close together; titles, descriptions, and anchor text represent at least part of the target page; and that anchor text may be a useful discriminator among unseen child pages. These results show the foundations necessary for the success of many web systems, including search engines, focused crawlers, linkage analyzers, and intelligent web agents.

This paper describes several known and some new methods for feature subset selection on large text data. Experimental comparison given on real-world data collected from Web users shows that characteristics of the problem domain and machine learning algorithm should be considered when feature scoring measure is selected. Our problem domain consists of hyperlinks given in a form of small-documents represented with word vectors. In our learning experiments naive Bayesian classifier was used on text data. The best performance was achieved by the feature selection methods based on the feature scoring measure called Odds ratio that is known from information retrieval.

For intelligent interfaces attempting to learn a user’s interests, the cost of obtaining labeled training instances is prohibitive because the user must directly label each training instance, and few users are willing to do so. We present an approach that circumvents the need for human-labeled pages. Instead, we learn “surrogate ” tasks where the desired output is easily measured, such as the number of hyperlinks clicked on a page or the amount of scrolling performed. Our assumption is that these outputs will highly correlate with the user’s interests. In other words, by unobtrusively “observing ” the user’s behavior we are able to learn functions of value. For example, an intelligent browser could silently observe the user’s browsing behavior during the day, then use these training examples to learn such functions and gather, during the middle of the night, pages that are likely to be of interest to the user. Previous work has focused on learning a user profile by passively observing the hyperlinks clicked on and those passed over. We extend this approach by measuring user mouse and scrolling activity in addition to user browsing activity. We present empirical results that demonstrate our agent can accurately predict some easily measured aspects of one’s use of his or her browser.

Tools for filtering the World Wide Web exist, but they are hampered by the difficulty of capturing user preferences in such a dynamic environment. We explore the acquisition of user profiles by unobtrusive monitoring of browsing behaviour and application of supervised machine-learning techniques coupled with an ontological representation to extract user preferences. A multi-class approach to paper classification is used, allowing the paper topic taxonomy to be utilised during profile construction. The Quickstep recommender system is presented and two empirical studies evaluate it in a real work setting, measuring the effectiveness of using a hierarchical topic ontology compared with an extendable flat list.

hours in the day anymore, are there? Everybody seems to have a stack of things waiting that just had to be done yesterday, and the problem only gets worse, not better. Information overload: it’s the bane of our times. This impression seems so realistic that any help in handling at least some simple tasks is usually appreciated. The Internet’s recent ascendancy—not only in the research community but also in many areas of everyday life—is a major culprit. No longer confined to providing researchers access to data, the Internet is often the source of choice for information about everything from what’s happening around the world, to where to get the best airline tickets, to how to cook a particular dish, to where to find the best hiking trails. It’s wonderful, but it also causes headaches. The impact on computer systems is particularly pronounced. When different people come together without centralized rules and guidance, many creative and pleasant, as well as some less pleasant, aspects emerge. So, when putting information on the World Wide Web, each of us can decide what to put there and how to organize it. The result is a distributed, world-wide-accessible information source that contains nonhomogeneous data organized according to different human asso-

Recently, Artificial Intelligence techniques have proved useful in helping users to handle the large amount of information on the Internet. The idea of personalized search engines, intelligent software agents, and recommender systems has been widely accepted among users who require assistance in searching, sorting, classifying, filtering and sharing this vast quantity of information. In this paper, we present a state-of-the-art taxonomy of intelligent recommender agents on the Internet. We have analyzed 37 different systems and their references and have sorted them into a list of 8 basic dimensions. These dimensions are then used to establish a taxonomy under which the systems analyzed are classified. Finally, we conclude this paper with a cross-dimensional analysis with the aim of providing a starting point for researchers to construct their own recommender system.

. The paper describes an approach to automatic Web-page classification based on the Yahoo hierarchy. Machine learning techniques developed for learning on text data are used here on the hierarchical classification structure. The high number of features is reduced by taking into account the hierarchical structure and using feature subset selection based on the method known from information retrieval. Documents are represented as feature-vectors that include n-grams instead of including only single words (unigrams) as commonly used when learning on text data. Based on the hierarchical structure the problem is divided into subproblems, each representing one on the categories included in the Yahoo hierarchy. The result of learning is a set of independent classifiers, each used to predict the probability that a new example is a member of the corresponding category. Experimental evaluation on real-world data shows that the proposed approach gives good results. For more than a half of testing...

Recommender systems learn about user preferences over time, automatically finding things of similar interest. This reduces the burden of creating explicit queries. Recommender systems do, however, suffer from cold-start problems where no initial information is available early on upon which to base recommendations. Semantic

In recent years, many systems and approaches for recommending information, products or other objects have been developed. In these systems, often machine learning methods that need training input to acquire a user interest profile are used. Such methods typically need positive and negative evidence of the user’s interests. To obtain both kinds of evidence, many systems make users rate relevant objects explicitly. Others merely observe the user’s behavior, which fairly obviously yields positive evidence; in order to be able to apply the standard learning methods, these systems mostly use heuristics that attempt to find also negative evidence in observed behavior. In this paper, we present several approaches to learning interest profiles from positive evidence only, as it is contained in observed user behavior. Thus, both the problem of interrupting the user for ratings and the problem of somewhat artificially determining negative evidence are avoided. The learning approaches were developed and tested in the context of the Web-based ELFI information system. It is in real use by more than 1000 people. We give a brief sketch of ELFI and describe the experiments we made based on ELFI usage logs to evaluate the different proposed methods.