D. Maltz and K. Ehrlich. Pointing the Way: Active Collaborative Filtering. In Proc. of CHI-95, pages 202--209, Denver, CO, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....and recommendation, to add the reader s personal opinion and viewpoint to the item. Categorise information, either manually, by using existing topic hierarchies like Syndic8, or automatically, using a technique such as Hyperspace Analog to Language (HAL) 9] 16] or collaborative filtering [10], 13] Social navigation techniques [5] to maintain historical data about information use. In general, our aim in the managing phase is to give users abstractions to allow them to interactively create and express connections both between information and between information and people where ....

Maltz, D., and K. Ehrlich. 1995. Pointing the way: Active collaborative filtering. Paper read at ACM Conference of Human Factors in Computing Systems (CHI'95), at Denver, CO.

.... NuggetMine has more in common with recommender systems, which use input from group members to make recommendations for other group members [21] Directly recommending items to other members of the group (as NuggetMine users do when they submit nuggets) is most like active collaborative filtering [17]. PILOT STUDY We deployed NuggetMine to a research lab at the University of Minnesota to qualitatively assess its benefits and weaknesses. There are six people in the lab, with some undergraduate and some graduate students. All lab members had met the other lab members in person; both authors ....

Maltz, D., Ehrlich, K. (1995). Pointing the Way: Active Collaborative Filtering. In Proceedings of CHI '95, pp. 202-209.

....track users as their interests change. However, such systems require an initial learning phase where behaviour information is built up to form an user profile. During this initial learning phase performance is often poor due to the lack of user information; this is known as the cold start problem [17]. There has been increasing interest in developing and using tools for creating annotated content and making it available over the semantic web. Ontologies are one such tool, used to maintain and provide access to specific knowledge repositories. Such sources could complement the behavioral ....

....content based recommendation is shared, allowing collaborative recommendation as well. We use the Quickstep [18] hybrid recommender system in this paper to recommend on line research 2. 1 The Cold start Problem One difficult problem commonly faced by recommender systems is the cold start problem [17], where recommendations are required for new items or users for whom little or no information has yet been acquired. Poor performance resulting from a coldstart can deter user uptake of a recommender system. This effect is thus self destructive, since the recommender never achieves good ....

Maltz, D. Ehrlich, E. Pointing the way: Active collaborative filtering, CHI'95 Human Factors in Computing Systems, 1995

....3: Example of an inference network consisting of 4 layers [TC92, p. 282] d = document nodes, r = concept representation nodes, q = query representation nodes, I = Interest node. 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 1 or L probabilities for node to become false probabilities for node to become true w 000 w 001 w 010 w 011 w 100 w 101 w 110 w 111 truth values of the 3 parent nodes 1 0 Figure 4: Link matrix implementing a weighted or connective. Values in the matrix are probabilities that the node takes the state determined by the row if parent nodes have the states determined by the ....

....held in the heads of the users. Users may know other users personally or from messages that these users originated or annotated. Active CF may be classified as either push active CF or pull active CF, depending on whether recommenders selects recipients, or recipients select recommenders [Mal94, ME95] In the simplest case, push) active CF may be performed by simply forwarding objects to other users, that one expects to be interested in these objects. Consequently, most electronic messaging systems may be considered as rudimentary active CF systems [MGT 87] Malone suggested allowing users ....

[Article contains additional citation context not shown here]

D. Maltz and K. Ehrlich. Pointing the way: active collaborative filtering. In Proceedings of the

....are also assumed to have more discriminatory value than an automated evaluation. The goal of social resource discovery systems is to aggregate and share the fruits of individual activity and knowledge. Relatively few social resource discovery systems currently exist. The Pointers system [3] facilitates the distribution of links to resources with accompanying context. While the benefits to a pointer s recipient seem clear, the system relies on a provider s desire to be helpful that may not always exist. Our work is closer in emphasis to Ringo [6] and GroupLens [5] GroupLens uses ....

Maltz, D., and K. Ehrlich. Pointing The Way: Active Collaborative Filtering. CHI 95: 202-209.

....PHOAKS: 60 March 1997 Vol. 40, No. 3 COMMUNICATIONS OF THE ACM the same types of benefits. In the case of ratingsbased systems, for example, everyone rates objects of interest. Yet there is evidence that people naturally prefer to play distinct producer consumer roles in the information ecology [2]; in particular, only a minority of people expend the effort of judging information and volunteering their opinions to others. Independently, we have observed such role specialization in Netnews; authors volunteer long lists of recommended Web resources at a stable, but low, rate. PHOAKS assumes ....

Maltz, D., and Ehrlich, K. Pointing the way: Active collaborative filtering. In Proceedings of the ACM Conference on Human Factors in Computing Systems, CHI'95 (Denver, May 7--11). ACM Press, New York., N.Y., pp. 202--209.

.... [Hill and Terveen, 1996] The particular quality of usage data that distinguishes it from evaluative data (such as Seals Of APproval [Rscheisen, Morgensen and Winograd, 1995] ratings [Allen, 1990; Resnick et al., 1994; Shardanand and Maes, 1995] and new associations [Bush, 1945; Kantor, 1993; Maltz and Erlich, 1995]) is that the perceived input cost to the recommender is zero. For example, users of the World Wide Web are often surprised at the level of detail a Web server can record about their activities without any awareness on their part that their usage was even being logged. Although there are costs ....

Maltz, D. and Erlich, K. (1995), Pointing the way: active collaborative filtering, Proceedings of the Conference on Human Factors in Computing Systems. CHI'95, Denver, CO, ACM, 202-9.

....in a database, and could be retrieved based not only on their content, but also on the opinions of others. For example, one could retrieve documents rated highly by a particular person or persons, or could retrieve documents whose annotations contained particular keywords. Maltz and Ehrlich [30] further developed this approach. They observed existing practice within organizations and noticed that a few individuals always played a very active role in making recommendations. They built a system designed expressly to support the two distinct roles of recommendation producer and user (or ....

Maltz, D. and Ehrlich, K. Pointing the Way: Active Collaborative Filtering, in Proceedings of CHI'95 (Denver CO, May 1995), ACM Press, 202-209.

....provider and recommendation recipient are specialized and different. PHOAKS reuses recommendations from existing online conversations. This reuse requires no extra work from providers, which distinguishes PHOAKS from collaborative filtering systems that make recommending an explicit task [5, 12] and from many group memory systems [1, 2, 16] Nor does PHOAKS require judgments of information quality from users, which is another difference from ratingsbased systems. Reusing information for new purposes forces either decontextualization or recontextualization of the information. PHOAKS ....

Maltz, D., and Ehrlich, K. Pointing the way: active collaborative filtering, in Proceedings of CHI'95 (Denver CO, May 1995), ACM Press, 202-209.

....enriched objects is that they wear out like real objects, and that visualizing usage of information highlights frequently used and therefore more promising objects. For examples of recent work on group memories and collaborative navigation that often make use of concepts like read wear, see [20, 31, 35]. Information richness in systems that support collaboration and interaction between users is an important ingredient to provide what we call social navigation [16] Social navigation is a behavior where users of information systems freely share pointers to information, and help out other users ....

Maltz, D. and Ehrlich, K. Pointing The Way: Active Collaborative Filtering, CHI'95, ACM Press, Denver, CO, 1995, pp. 202-209.

....neighborhood based predictive algorithms, implement a single collaborative model that can be described as Birds of a Feather Flock Together (BOF) These systems tightly couple the architecture and the collaborative model. In the case of PHOAKS, Referral Web, and active collaborative filtering [12], different algorithms and different collaborative models were chosen because the BOF model and clustering algorithms did not effectively fit the recommendation situation. In expertise recommendation, a BOF model that clusters profiles may not effectively distinguish individuals who have expertise ....

Maltz, D. and Ehrlich, K., Pointing The Way: Active Collaborative Filtering. CHI '95, 202 - 209.

.... Document can then be treated as vectors in a high dimensional semantic space and similarity is based on the angle between such vectors (Salton 1989) or advanced pattern matching algorithms (e.g. Autonomy 1998) The third approach was chosen for this study since it avoids the cold start problem (Maltz Ehrlich 1995), i.e. it does not require a critical mass of users to pay off but provides benefit also for the very first user. Finally, recommender systems can be combined with agent technology. If included, such technology enables a certain degree of autonomy that allows continuos monitoring without user ....

Maltz, D. and Ehrlich, K. (1995). "Pointing the way: Active collaborative filtering", In Proceedings of CHI '95, ACM Press: Denver, CO.

....jokes to the user and continues to collect ratings on each recommended joke. The communication between the interface and the server scripts takes place through a CGI script written in C. Figure 3 illustrates the interface. All collaborative filtering systems experience the cold start problem [22]. One needs ratings to predict ratings. To address this, we started with a simple website to collect joke ratings. We chose the initial set of 40 jokes from friends and newsgroups, doing our best to avoid highly offensive jokes. We then asked 80 friends and students to rate all 40 jokes by ....

D.A. Maltz and K. Ehlrich. Pointing the way: Active collaborative filtering. In Chi'95 Proceedings Papers, 1995.

....Frequency) 71, 72] Latent Semantic Indexing [23, 7] and edit distances [47, 90] User Based Classifiers Rather than build artificial feature extractors directly, humans themselves can judge and classify the records in the database. This is often used in collaborative filtering applications [52, 67, 34] where a user specifies a small set of positive examples of the desirable class, upon which retrieval of other potentially desirable samples is based. The basic mechanism behind a system such as this is that the classification into positive (specified) samples and everything else (negative ....

D. Maltz and K. Ehrlich. Pointing the way: Active collaborative filtering. In Proceedings of the 1995 ACM Conference on Computer Human Interaction, pages 201--209, 1995.

....service. Personalising content necessarily means understanding content consumers (users) Recent research has been devoted to Artificial Intelligence (AI) techniques for automatically modelling and profiling user needs (see example [3,4] and for matching such models with relevant content items [2, 7]. PTV is one such system. The basic idea behind PTV is the online personalised TV guide . That is, PTV is a television guide, listing programme viewing details just like any other guide, but with one important difference, the listed programmes are carefully selected to match the viewing ....

.... approach is the feature based approach epitomised by information retrieval (IR) and case based reasoning research [5; 8] A more recent strategy, which is gaining popularity in Web based applications, is the feature less approach, epitomised by techniques such as collaborative filtering [2,6,7]. With the former technique, content items and user queries (or profiles) are described using a set of carefully chosen descriptive features. Content recommendation is all about matching content items against user queries (or profiles) in terms of their descriptive features, the best matching ....

Maltz D., Ehrlich K., "Pointing the Way: Active Collaborative Filtering", Proceedings of the ACM Conference on Human Factors in Computing Systems, CHI '95, ACM Press, New York, N.Y., pp. 202-209, 1995.

....[3] that is, the idea that online users can be automatically clustered into groups with similar tastes and preferences. Closely tied with this concept is the technique of collaborative filtering, which is an approach to information filtering and recommendation that uses the virtual community idea [1, 4, 5, 7]. Very simply, information for a particular user is filtered or recommended according to that user s membership in a particular community. In other words, a target user will be recommended information items that other members of her virtual community have liked, and similarly, information items ....

D. Maltz, K. Ehrlich (1995) Pointing The Way: Active Collaborative Filtering. Proceedings of CHI 95, Denver, Colarado, USA.

....Collaborative recommendation methods such as automated collaborated filtering are an alternative to case based techniques. Instead of recommending new programmes that are similar to the ones that the user has liked in the past, we recommend programmes that other similar users have liked ([1, 3, 4, 8 ,9, 10]) Rather than compute the similarity between items, we compute the similarity between users, or more precisely the similarity between user profiles. Note that we have opted for a lazy approach to collaborative filtering rather than the more traditional eager approach where the user base is ....

Maltz D., Ehrlich K.: Pointing the Way: Active Collaborative Filtering. In: Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI '95) ACM Press, New York, N.Y., (1995). 202-209

....going on in an organisation. For example, if a document is never accessed it is likely that its content is of little value for a majority of the personnel. Interesting documents are on the other hand likely to be found and accessed, partly because they are often passed along to colleagues (see, Maltz and Ehrlich 1995). This paper describes a system, named IntraNews, that help people to find out where interesting things are happening on intranets. The system is a news recommending system based on collaborative filtering (for overview see, Resnick and Varian 1997) The design of IntraNews was informed by ....

....provide users with new resources through social filtering. Siteseer (Rucker and Polanco 1997) does this by matching the users bookmark files and Yenta (Foner 1997) extracts information from many sources, e.g. Web documents and emails. This approach needs a critical mass of users to pay off (see Maltz and Ehrlich s (1995) discussion of the cold start problem ) in contrast to our system, where even one single user could benefit. Avery and Zeckhauser (1997) explored the problems that content owners may manipulate recommender systems by recommending them self very strong. This is hardly going to be a problem here, ....

Maltz, D. and K. Ehrlich (1995) "Pointing the way: Active collaborative filtering," In Proceedings of the ACM CHI '95 conference, Denver, CO, 202-209.

....are also assumed to have more discriminatory value than an automated evaluation. The goal of social resource discovery systems is to aggregate and share the fruits of individual activity and knowledge. Relatively few social resource discovery systems currently exist. The Pointers system [3] facilitates the distribution of links to resources with accompanying context. While the benefits to a pointer s recipient seem clear, the system relies on a provider s desire to be helpful that may not always exist. Our work is closer in emphasis to Ringo [6] and GroupLens [5] GroupLens uses ....

Maltz, D., and K. Ehrlich. Pointing The Way: Active Collaborative Filtering. CHI 95: 202-209.

....are only as good as the information they contain. Systems that make suggestions without an explicit representation of the content are one possible solution this problem. Traditionally, recommender systems are an approach to solving problems of information overload. Social filtering systems [3, 4], collaborative filtering systems [2] and rating systems can all be considered recommender systems [6] These systems highlight or deliver relevant items to the user. They maintain a profile of each user based on prior system utilization or information which the user supplies. In all of the ....

Maltz, D. and K. Ehrlich. Pointing The Way: Active Collaborative Filtering. Proceedings of CHI '95, 1995.

....GroupLens support a single keystroke rating input (or, when possible, replacing an existing keystroke) since users typically spend very little time or attention on any particular article. Other research has shown that more extensive textual ratings can be effective in close knit communities [2, 4]. The requirement that GroupLens provide predictions of the rating the system expects the user will give each article, rather than only winnowing down the list of articles. We consider it very important to provide advice rather than exercise censorship. The pilot study, successful yet ....

....though we do not have the resources to serve that large a population and data set except perhaps with an overall average prediction rather than personalized predictions. Usenet presents a different set of challenges to collaborative filtering than domains such as music [12] or movies [4] where new items are relatively infrequent and lifetimes are relatively long. In addition to addressing critical performance issues, the GroupLens system continues to address several key problems involving ratings sparsity and start up usage by applying techniques including partitioning the ....

Maltz, D. and Ehrlich, K. Pointing the way: Active collaborative filtering. In Proceedings of the 1995 ACM Conference on Human Factors in Computing Systems. (1995). ACM, New York.

No context found.

D. Maltz and K. Ehrlich. Pointing the Way: Active Collaborative Filtering. In Proc. of CHI-95, pages 202--209, Denver, CO, 1995.

No context found.

Maltz, D., Ehrlich, K. (1995): Pointing the way: active collaborative filtering. Proceedings of the Conference on Computer-Human Interaction. pp. 202--209

No context found.

Maltz, D. and Ehrlich, K. Pointing the Way: Active Collaborative Filtering, in Proceedings of CHI'95 (Denver CO, May 1995), ACM Press, 202-209.

No context found.

Maltz D., Ehrlich K. (1995) Pointing the Way: Active Collaborative Filtering. Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI '95) ACM Press, New York, N.Y., 202-209

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC