Muh-Chyun Tang

Query length in interactive information retrieval

Query length in best-match information retrieval (IR) systems is well known to be positively related to effectiveness in the IR task, when measured in experimental, non-interactive environments. However, in operational, interactive IR systems, query length is quite typically very short, on the order of two to three words. We report on a study which tested the effectiveness of a particular query elicitation technique in increasing initial searcher query length, and which tested the effectiveness of queries elicited using this technique, and the relationship in general between query length and search effectiveness in interactive IR. Results show that the specific technique results in longer queries than a standard query elicitation technique, that this technique is indeed usable, that the technique results in increased user satisfaction with the search, and that query length is positively correlated with user satisfaction with the search.

Exploring alternative metrics of scholarly performance in the social sciences and humanities in Taiwan

Research output and impact metrics derived from commercial citation databases such as Web of Science and Scopus have become the de facto indicators of scholarly performance across different disciplines and regions. However, it has been pointed out that the existing metrics are largely inadequate to reflect scholars’ overall peer-mediated performance, especially in the social sciences and humanities (SSH) where publication channels are more diverse. In this paper alternative metrics exploring a variety of formal and informal communication channels were proposed, with the aim of better reflecting SSH scholarship. Data for a group of SSH scholars in Taiwan on these metrics were collected. Principal component analysis revealed four underlying dimensions represented by the 18 metrics. Multiple-regression analyses were then performed to examine how well each of these dimensions predicted the academic standing of the scholars, measured by the number of public grants awarded and prestigious research awards received. Differences in the significance of the predictors were found between the social sciences and humanities. The results suggest the need to consider disciplinary differences when evaluating scholarly performance.

Representational practices in digital museums: A case study of the National Digital Museum Project of Taiwan

The study explores the semiotic aspects of the migration of museum artifacts to the digital domain. Seeing digital museums from a semiotic perspective helps foreground the representational practices utilized and the “museum effects” they create in the digital environment. Three levels of representational practices were discussed: the representation of the original artifacts by their digital images, the representation of the national consciousness by the cultural elements highlighted by the collections, and the representational schemes that provide access to cultural materials. A case study of the theme projects under Taiwans National Digital Museum Project is presented with a view to examining current representational schemes for cultural materials. The relations among curatorial intention, the nature of the cultural image, and the representational schemes are explored. The implications of current practices on the dissemination of cultural materials online are discussed.

Inequality of publishing performance and international collaboration in physics

Using a database of 1.4 million papers indexed by Web of Science, we examined the global trends in publication inequality and international collaboration in physics. The publication output and citations received by authors hosted in each country were taken into account. Although inequality decreased over time, further progress toward equality has somewhat abated in recent years. The skewedness of the global distribution in publication output was shown to be correlated with article impact, that is, the inequality is more significant in articles of higher impact. It was also observed that, despite the trend toward more equalitarian distribution, scholarly participation in physics is still determined by a select group. Particularly noteworthy has been Chinas rapid growth in publication outputs and a gradual improvement in its impact. Finally, the data also suggested regional differences in scientific collaboration. A distinctively high concentration of transnational collaboration and publication performance was found among EU countries.

A longitudinal study of intellectual cohesion in digital humanities using bibliometric analyses

As digital humanities continues to expand and become more inclusive, little is known about the extent to which its knowledge is integrated. A bibliometric analysis of published literature in digital humanities was conducted to examine the degree of its intellectual cohesion over time (1989–2014). Co-authorship, article co-citation, and bibliographic coupling networks were generated so SNA based cohesion analysis can be applied. Modularity maximization partition was also performed to both co-citation and “author bibliographic coupling” networks to identify main research interests manifested in the literature. The results show that, as publications in digital humanities continue to grow, its diversity and coherence, two hallmarks of interdisciplinarity, have shown signs of becoming more robust. The co-author network, however, remained rather fragmented, with collaboration mainly limited by language and geographic boundaries. The domain specific practices in digital humanities that might contribute to such fragmentation was discussed.

Not All Books in the User Profile Are Created Equal: Measuring the Preference “Representativeness” of Books in aNobii Online Bookshelves

The study proposes a novel construct of “representativeness” that aims to measure the degree to which a book in the user’s online bookshelf is able to represent his/her reading preference, based on the assumption that not all books are equally important when it comes to constructing individual users’ preference profiles. Thirty-five online bookshelf aNobii users were recruited, who were asked to perform a judgment task involving evaluating the degree of “representativeness” and “involvement” of 10 books self-selected from their bookshelves. The results show that there is a high correlation between “representativeness’ and “involvement”, a well-established construct in marketing. Book similarity networks for every participants was generated based on book co-ownership data extracted from aNobii. Two social network analytical (SNA) metrics: coreness and connectivity, were then applied to measure a book’s “representativeness” relative to the individual bookshelves. Results show that there were significant correlations between the SNA metrics and the user’s self-assessed “representativeness” and “involvement” of the books. Furthermore, it was found the correlations were stronger among bookshelves owned by users who have low reading diversity.

Exploring the Impact of Users’ Preference Diversity on Recommender System Performance

Recommender systems present an effective alternative to subject access in the domain of reading for leisure. They are particular valuable for the discovery of novel and serendipitous finds. In the recommender system evaluation literature, a trade-off has been recognized between accuracy and surprise/non-obviousness, and more recently, that between accuracy and diversity of the recommendation set. It is argued in this paper that a proper balance between accuracy and diversity might lie in users’ “preference diversity,” a construct we propose to represent how wide a user’s reading interests are. Users with more diverse interests might appreciate more novel and diverse set of recommendation. Drawing from marketing literature and our own empirical studies, this paper discusses how the motivation for diversity might influence users’ preference for cultural goods such as books and music, and more specifically their responses to recommendations. Future study is needed to further examine the relationship between user preference diversity and proper degree of diversity of the recommendation set.

Evaluating a metadata-based term suggestion interface for PubMed with real users with real requests

This paper reports results of an evaluation study of MAP (Multi-faceted Access to PubMed), a metadata induced query suggestion interface for PubMed bibliographic search. A novel evaluation methodology was used to address the challenges involved in evaluating an IIR (Interactive Information Retrieval) system such as the MAP interface. The most significant aspect of this methodology is that, instead of using assigned tasks common in traditional IR evaluation, it asks real users with real search requests to search with real systems in an experimental setting. Several performance measures were created based on which comparisons were made between MAP and PubMed baseline. MAP was shown to perform better in several of these measures, especially when the search requests had not been attempted before. The finding pointed to search characteristics as an important intervening variable in IIR evaluation. The advantages of and potential threats to our methodology were also discussed.

Multiple access to PubMed: A proposal of utilizing MeSH as a term suggestion tool for PubMed bibliographic search

Origin of the Problem One of the critical problems faced by the user searching for a heterogeneous and massive bibliographic database such as PubMed is the difficulty of managing the returned results. Several system features of PubMed (e.g. default explode function and free-text indexing in title and abstract fields) that aim at facilitating end-user search tend to increase indexing exhaustivity and therefore favor search recall at the expense of precision. Faced with the unmanageable amount of returned results and without an efficient tool to systematically narrow down their searches, users are often left with few options but to hastily browse the first few returned pages. The consequence of information overload creates at least two barriers to a successful user-system communication. Firstly, there in no telling whether there might be documents relevant to users needs buried deep down in the returned set that never get the chance of being viewed. Secondly, the skimming of the surface of the returned set also gives incomplete feedback for the judgment of users query performance. We propose an interface that makes use the existing structure of MeSH (Medical Subject Headings) to help users manage search results. The proposed solution: using MeSH as a term suggestion tool In an automatic indexing environment where the indexing exhaustivity is quite high, short queries usually produce unmanageably large returned sets. The lack of precision of the search can be alleviated, however, by eliciting longer queries from the user. To elicit more terms from users, we propose a term suggestion feature that utilizes the classification structure of MeSH. The tool involves extracting MeSH terms present in the returned set after the user submit her/his query. These terms are considered conceptually associated with the original query due to their co-occurrences in the retrieved records. Extracted terms presentation It is crucial to organize and present the extracted MeSH terms in a manner that makes efficient browsing and selection possible. We plan to experiment and compare two ways of presenting the MeSH terms to the user. The baseline system will simply rank the terms based on their frequency in the original returned set. The other presentation method will organize the extracted terms using the 15 top-level semantic categories of MeSH. Terms co-occur with the initial query will be mapped against the MeSH tree with the non-occurring terms left out. Thus a filtered MeSH tree that includes only the extracted terms will be generated dynamically each time a query is submitted. The user will be given the option of browsing and selecting terms from the post-querying MeSH tree.