Search

Purpose: One typical document retrieval method is to use to co-citation. The method is based on the premise that the degree of similarity among co-cited papers is equal in a particular paper. The degree is calculated with binary values: "co-cited" or "not-cited". To improve upon this method, the author proposes a multivalued co-citation measure based on semantic distance between co-cited papers. Methods: To determine the distance between citations, the author measured two machine parseable relationships (location and citing words) between places where papers are cited. In order to evaluate the proposed method, we identified two categories of co-citation: a group with strong relationships indicating "enumerated co-citation" (papers cited within one statement) and a group with weak relationships showing "non enumerated co-citation". Similarities within each group were calculated and compared using the CiteSeer dataset and 6 major similarity indicators. Results: All of the similarity indicators showed that degree of "enumerated co-citation" is higher than "non enumerated co-citation". Consequently, it became clear that the proposed co-citation measure can be used to distinguish the strength of co-citation more precisely and the it can be applied to large-scale document collections. [ABSTRACT FROM AUTHOR]

In this paper, we propose a method to find query suggestions of a verbal query, which contains a verb in the query, from the Web. People sometimes cannot obtain appropriate search results even if they consider they have formulated a query that clearly describes their search intents. The idea of the proposed method is to find the relationship between verb and noun in the query, and mine the appropriate representation of the verb based on the relationship. The proposed method estimates the relationship between verb and noun based on particles between them. Based on the estimated relationship, we then obtain candidates of the verb in the query by using either the Web search results or the case frame. Next, we compute the effectiveness of the candidates by considering the similarity between a candidate and the verb and the co-occurrence between the candidates and the noun, and finally rank the candidates to generate queries. To investigate the effectiveness of our proposed method, we conducted the experiment by comparing with the query suggestions of a commercial search engine as our baseline. The experimental result of 20 queries showed that our proposed method, which finds candidates from the Web search results, outperformed the baseline method in terms of AvgRelNum, which measures the the number of relevant pages obtained by the generated query that can retrieve a relevant page, and achieved the similar performance in terms of Contain@10 and MRR@10. [ABSTRACT FROM AUTHOR]

In recall-oriented tasks, where collecting extensive information from different aspects of a topic is required, searchers often have difficulty formulating queries to explore diverse aspects and deciding when to stop searching. With the goal of helping searchers discover unexplored aspects and find the appropriate timing for search stopping in recall-oriented tasks, this paper proposes a query suggestion interface displaying the amount of missed information (i.e., information that a user potentially misses collecting from search results) for individual queries. We define the amount of missed information for a query as the additional gain that can be obtained from unclicked search results of the query, where gain is formalized as a set-wise metric based on aspect importance, aspect novelty, and per-aspect document relevance and is estimated by using a state-of-the-art algorithm for subtopic mining and search result diversification. Results of a user study involving 24 participants showed that the proposed interface had the following advantages when the gain estimation algorithm worked reasonably: (1) users of our interface stopped examining search results after collecting a greater amount of relevant information; (2) they issued queries whose search results contained more missed information; (3) they obtained higher gain, particularly at the late stage of their sessions; and (4) they obtained higher gain per unit time. These results suggest that the simple query visualization helps make the search process of recall-oriented tasks more efficient, unless inaccurate estimates of missed information are displayed to searchers. [ABSTRACT FROM AUTHOR]

As open-access policies gain acceptance, an increasing number of researchers are contributing their papers to publicly accessible web sites (i.e. self-archiving). Theoretically, these papers are accessible from standard search engines, but they tend to be obscured by other contents on the web. The purpose of this research is to develop a system that can automatically detect academic articles and/or quasi-academic articles on the web. This paper describes experiments that were conducted on the performance of various classifiers and the results are compared in terms of precision, recall, and F-measure. The classifiers use attributes such as terms in PDF files and empirical rules. The results suggest the efficiency of a ranked output system which has several phases to identify academic articles. [ABSTRACT FROM AUTHOR]

The World Wide Web has had an unprecedented impact on the creation and utilization of information. As the amount of information available on the Web has continued to increase, it has become more difficult to decide whether certain information is reliable or not. This paper analyzes the information seeking behavior of Web users and examines their information evaluation processes. It is based on experimental research on students from a two-year college program. The results of the study are intended to be used by libraries to enhance user education. The paper reviews previous studies, examining studies on Web searching tendencies and the evaluation of Web resources in accordance with their methodologies. We discuss the simplicity of Web searching patterns and examine the factors that influence information evaluation. We point out the limitations of the research methods that have been used in the past, and describe the reasons for adopting the protocol analysis method in this investigation. We used an observational method and protocol analysis to investigate the students' behavior as they used search engines and the OP AC (Online Public Access Catalogue) to find information. We divided the students into two groups: students who have used the Web for more than five years and students who have used the Web for less than two years. In order to provide a basis for comparison, we repeated this research on university students. We came to the following conclusions. Web searching is a repeated behavior involving the repetition of a simple, regular pattern. The students were able to filter out unnecessary information quickly as they evaluated search results. Experience affected information seeking performance and evaluation, and a more pronounced effect was noted with students who had a longer history of experience with the Web. Students had a tendency to evaluate Web resources based on visual factors and personal experiences, but they lacked skills in determining the quality of [ABSTRACT FROM AUTHOR]

The document clustering technique is widely recognized as a useful tool for information retrieval, organizing web documents, text mining and so on. The purpose of this paper is to review various document clustering techniques, and to discuss research issues for enhancing effectiveness or efficiency of the clustering methods. We explore extensive literature on non-hierarchical methods (single-pass methods), hierarchical methods (single-link, complete-link, etc.), dimensional reduction methods (LSI, principal component analysis, etc.), probabilistic methods, data mining techniques, and so on. In particular, this paper focuses on typical techniques, such as the k-means algorithm, the leader-follower algorithm, self-organizing map (SOM). single- or complete-link methods, bisecting k-means methods, latent semantic indexing (LSI), Gaussian-Mixture model and so on. After reviewing the techniques and algorithms, we discuss research issues on document clustering: computational complexity, feature extraction (selection of words), methods for defining term weights and similarity, and evaluation of results. [ABSTRACT FROM AUTHOR]

Purpose: In recent years, "a work" is attracting attention as an entity for information retrieval. This paper investigates associations of works which are called "bibliographic family" represented in the Japanese library catalog. Methods: To investigate bibliographic families in the Japanese library catalog, a survey was conducted using a random sample from J-BISC. Research questions were: 1) what proportion of works in the Japanese library catalog have derivative works, 2) how populous each bibliographic family is, 3) how frequently each type of derivative relationship appears, and 4) whether there are unique characteristics of bibliographic relationships in Japan compared with studies from other countries. From J-BISC, 669 works that had been recorded up to January 2005 were extracted as the sample. To search for members of bibliographic families, NDL-OPAC and Webcat Plus was used. To investigate the characteristics of bibliographic families of works in the Japanese library catalog, the results of previous studies were compared and examined. Results: The survey results show that 1) 2 5.9% of works in the sample have derivative works, 2) the average size of bibliographic families in the sample is 1.80, and 3) successive relationships account for 63.6% of all types of derivative relationships. Based on these results, networks of bibliographic families really exist, and so it is valid to use the links to improve the library catalog. From a comparison with five previous studies in the U.S. and Slovenia, the phenomenon of works which are called bibliographic families is free from locality and language, and shows "universality." [ABSTRACT FROM AUTHOR]

Sanda Erdelez's information behavior model of information encountering does not satisfactorily explain the relationship between active and passive information behaviors, and the types and definitions of unit behaviors which together constitute the process of information encountering. This paper examines Erdelez's model with regards to these two issues proposes a revised model, and discusses the validity of the model. The revised model has three main characteristics (1) The information encountering process is independent from the task which generates it, and the user does not necessarily go back to the original task (2) The information encountering process consists of the following eight unit behaviors: "noticing", "stopping", "examining", "use", "sharing", "storing", "returning", and "ending" (3) The unit behaviors are inter-related. In order to discuss the validity of the revised model, a series of qualitative interviews were conducted on information encountering on the Web. The results favor the validity of the model and show some new aspects of information encountering, some of which are presumably peculiar to the Web. These may contribute to further revisions of the model. [ABSTRACT FROM AUTHOR]