Jack Duffy

Effect of task on time spent reading as an implicit measure of interest

Information Filtering systems learn user preferences either through explicit or implicit feedback. However, requiring users to explicitly rate items as part of the interface interaction can place a large burden on the user. Implicit feedback removes the burden of explicit user ratings by transparently monitoring user behavior such as time spent reading, mouse movements and scrolling behavior. Previous research has shown that task may have an impact on the effectiveness of some implicit measures. In this work we report both qualitative and quantitative results of an initial study examining the relationship between user time spent reading and relevance for three web search tasks: relevance judgment, simple question answering and complex question answering. This study indicates that the usefulness of time spent as a measure of user interest is related to task and is more useful for more complex web search tasks. Future directions for this research are presented.

Using large tables on small display devices

The next evolutionary step in wireless Internet information management is to provide support for tasks, which may be collaborative and may include multiple target devices, from desktop to handheld. This means that the information architecture supports the processes of the task, recognizes group interaction, and lets users migrate seamlessly among internet-compatible devices without losing the thread of the session. If users are free to migrate amongst devices during the course of a session then intelligent transformation of data is required to exploit the screen size and input characteristics of the target appliance with minimal loss of task effectiveness.In this paper we first review general characteristics related to the performance of users on small screens and then examine the navigation of full tables on small screens for users in multidevice scenarios. We examine the methodologies available for access to full tables in environments where the full table cannot be viewed in its entirety. In particular, we examine the situation where users are collaborating across platform and referring to the same table of data. We ask three basic questions: Does screen size affect the performance of table lookup tasks? Does a search function improve performance of table lookup based tasks on reduced screen sizes? Does including context information improve the performance of table lookup based tasks on reduced screen sizes? The answers to these questions are important as individual and intuitive responses are used by the designers of small screen interfaces for use with large tables of data. We report on the results of a user study that examines factors that may affect the use of large tables on small display devices. The use of large tables on small devices in their native state becomes important in at least two circumstances. First, when collaboration involves two or more users sharing a view of data when the individual screen sizes are different. Second, when the exact table structure replication may be critical as a user moves quickly from a larger to a smaller screen or back again mid-task. Performance is measured by both effectiveness, correctness of result, and efficiency, effort to reach a result.

An Examination of Genre Attributes for Web Page Classification

In this paper, we describe a set of experiments to examine the effect of various attributes of web genre on the automatic identification of the genre of web pages. Four different genres are used in the data set, namely, FAQ, News, E-Shopping and Personal Home Pages. The effects of the number of features used to represent the web pages (5, 20, or 100) as well as the types of attributes, <content, form, functionality>, singly and in various combinations are examined. The results indicate that fewer features produce better precision but more features produce better recall, and that attributes in combinations will always perform better than single attributes.

Web Page Transformation When Switching Devices

With network and small screen device improvements, such as wireless abilities, increased memory and CPU speeds, users are no longer limited by location when accessing on-line information. We are interested in studying the effect of users switching from a large screen device, such as a desktop or laptop to use the same web page on a small device, in this case a PDA (Personal Digital Assistant). We discuss three common transformation approaches for display of web pages on the small screen: Direct Migration, Linear and Overview. We introduce a new Overview method, called the Gateway, for use on the small screen that exploits a user’s familiarity of a web page. The users in an initial study prefer using the Gateway and Direct Migration approach for web pages previously used on the large screen, despite the common Linear approach used by many web sites.

An N-Gram Based Approach to Automatically Identifying Web Page Genre

The research reported in this paper is the first phase of a larger project on the automatic classification of web pages by their genres, using ngram representations of the web pages. In this study, the textual content of web pages is used to create feature sets consisting of the most frequent n-grams and their associated frequencies. We present three methods, each of which uses a distance measure to determine the dissimilarity between two feature sets. Each method forms a feature set for every web page in the test set, however the formation of feature sets from the training set differs between methods: we experiment using one feature set per web page, per genre, and a combination of genre-based feature sets supplemented by subgenre feature sets. We present results for a balanced corpus of seven genres (blog, eshop, FAQs, front page, listing, home page, and search page). Initial results are encouraging.

Applications of Artificial Intelligence Systems in the Analysis of Epidemiological Data

A brief review of the germane literature suggests that the use of artificial intelligence (AI) statistical algorithms in epidemiology has been limited. We discuss the advantages and disadvantages of using AI systems in large-scale sets of epidemiological data to extract inherent, formerly unidentified, and potentially valuable patterns that human-driven deductive models may miss.

Diversity and Connectance in an Industrial Context

The ecological metaphor of industrial ecology is a proven conceptual tool, having spawned an entire field of interdisciplinary research that explores the intimate linkages between industry and its underlying natural systems. Besides its name and a number of borrowed concepts, however, industrial ecology has no formal relationship with the ecological sciences. This study explores the potential for further interdisciplinary collaboration by testing whether some of the same quantitative analysis techniques used in community ecology research can have meaning in an industrial context. Specifically, we applied the ecological concepts of connectance and diversity to an analysis of Burnside Industrial Park in Halifax, Nova Scotia. Our results demonstrate that these ecological tools show promise for use in industrial ecology. We discuss the meaning of connectance and diversity concepts in an industrial context and suggest next steps for future studies. We hope that this research will help to lay the groundwork of an ecologically inspired tool kit for analyzing industrial ecosystems.

Comparing two one-handed access methods on a PDA

Users of mobile devices often need to use those devices in contexts which leave only one hand available for manipulating the device, such as holding another device or manual, walking or operating some machinery.In this paper we discuss the results of a comparison of the effectiveness, efficiency and preference users have for map navigation tasks on a PDA, where they are restricted to one handed use. One method uses a tilt sensor and touch screen and the other uses multifunction buttons and the touch screen.The results of this study indicate that neither method was significantly more effective (accurate), efficient, or preferred by the users for one handed manipulation of three maps. We did find indications, however, that the tilt method helped users create better cognitive overviews of the maps.

A Patient Profile Ontology in the Heterogeneous Domain of Complex and Chronic Health Conditions

There is growing interest in recent years in applying ontologies to represent disease concepts because they have the ability to depict the domain knowledge with a superior level of expressiveness and precision. Ontologies have been predominantly used to represent well-categorized disease concepts. However, there are challenges in representing the domain knowledge for heterogeneous and poorly categorized systems. In this study, a methodology to create an ontology to represent the domain knowledge for complex and chronic health conditions is explored. The domain of complex chronic conditions can be viewed not only as heterogeneous but also as dynamic with new knowledge continually evolving. The methodology includes the development of a controlled vocabulary to create the first layer of semantic interoperability. The controlled vocabulary is then converted into a patient profile ontology to add deeper semantics, conceptually and relationally in the heterogeneous domain knowledge.

Binary Cybergenre Classification Using Theoretic Feature Measures

In this study, we conducted an investigation on automatic genre classification for three common types of Web pages addressing the effect of three theoretic feature selection measures, a range of feature set size, and three machine classifiers on the accuracy of the Web page classification in the context of a set of controlled experiments. Our results are encouraging and we conclude that for binary classification tasks, at least for these Web page genres, it is possible to reach satisfying results with small content-based feature sets generated with a sound feature selection measure and furthermore there is no evidence of interaction between these feature selection measures and the machine classifiers used