Building and especially improving a classification kernel represents a challenging task. The works presented in this paper continue an already developed semi-supervised classification approach that aimed at labelling transcripts from educational videos. We questioned whether the size of the ground-truth data-set (Wikipedia articles) or the quality…

Introduction: The use of averaged topic-level scores can result in the loss of valuable data and can cause misinterpretation of the effectiveness of system performance. This study aims to use the scores of each document to evaluate document retrieval systems in a pairwise system evaluation. Method: The chosen evaluation metrics are document-level…

This paper discusses the provision of flexible querying facilities over heterogeneous data arising from lifelong learners' educational and work experiences. A key aim of such querying facilities is to allow learners to identify possible choices for their future learning and professional development by seeing what others have done. We motivate and…

This article compares three self-developed subject guides and one commercial online subject guide (LibGuides) to provide information to other institutions which have a plan or are in the process of developing a plan to implement similar services to support online, distance and traditional education. This is a comparison of the current subject…

Presents two partitioning methods for signature files and a method of fast weight computation to implement a document ranking strategy. Storage requirements, number of disk accesses, and processing time for these methods are compared with two other access methods. Merits of the partitioning methods are summarized. (20 references) (MES)

Differences in judged relevance of Educational Resources Information Center (ERIC) abstracts identified by two different information systems were investigated. A national semi-private educational service agency which utilizes mechanical searching of the research literature was one information system, while the other was a regional,…

The study compares the retrieval effectiveness of two alternative input and search systems in terms of such measures as recall, fallout, precision, and total retrieval. One system uses manually indexed document files searched by controlled vocabulary, while the other employs fulltext input using natural language searching. Both systems are applied…

Reviews past research on user-satisfaction and relevance information systems. Describes an experiment that compares user-satisfaction with retrieved items and the retrieval process with the relevance of the items in resolving user information needs. Results indicate a connection exists between relevance and user-satisfaction with the retrieved…

A task-oriented, comparative evaluation between four Web retrieval systems was performed; two using query-biased summarization, and two using the standard ranked titles/abstracts approach. Results indicate that query-biased summarization techniques appear to be more useful and effective in helping users gauge document relevance than the…

Integrates structured data and text using the unchanged, standard relational model. The performance of a parallel relational database system and the effect of query reduction on accuracy were tested, and it was found that queries can be reduced prior to their implementation without incurring significant loss in precision/recall. (AEF)

Compares two natural language processing search engines, WIN (WESTLAW Is Natural) and FREESTYLE, developed by LEXIS. Legal issues in natural language queries were presented to identical libraries in both systems. Results showed that the editorials enhanced relevance; a search would be more thorough using both databases; and if only one system were…

In assessing information retrieval systems, it is important to compare the number of retrieved relevant items to the total number of relevant items. A variant of the statistical "capture-recapture" method can be used in large collections to estimate total number of relevant documents, providing the model supporting such an analysis can…

Retrieval performance of the iso-content-based angle similarity measure within the angle, distance, conjunction, disjunction, and ellipse retrieval models is compared with retrieval performance of the distance similarity measure and the angle similarity measure. Results show the iso-content-based angle similarity measure achieves satisfactory…

Reports on comparative study of document classifications produced by use of single linkage, complete linkage, group average, and Ward clustering methods. Findings of work that compares use of clusters consisting of pairs of documents with conventional best match searches are also reported. Thirty-four references are provided. (EJS)

Presents criteria for an effective syntagmatic relator system and describes how conceptual syntagmatic relationships are expressed in natural language, commenting on the adequacy of such expressions for document retrieval systems. Only one type of expression, a gestalt structure, represented as a frame, conforms to the criteria. (Author/JKP)