Nigel Stanger

Context identification of sentences in related work sections using a conditional random field: towards intelligent digital libraries

Identification of contexts associated with sentences is becoming increasingly necessary for developing intelligent information retrieval systems. This article describes a supervised learning mechanism employing a conditional random field (CRF) for context identification and sentence classification. Specifically, we focus on sentences in related work sections in research articles. Based on a generic rhetorical pattern, a framework for modelling the sequential flow in these sections is proposed. Adopting a generalization strategy, each of these sentences is transformed into a set of features, which forms our dataset. We distinguish between two kinds of features for each of these sentences viz., citation features and sentence features. While an overall accuracy of 96.51% is achieved by using a combination of both citation and sentence features, the use of sentence features alone yields an accuracy of 93.22%. The results also show F-Scores ranging from 0.99 to 0.90 for various classes indicating the robustness of our application.

Ontology-based modelling of related work sections in research articles: using CRFs for developing semantic data based information retrieval systems

Research articles are an important form of scientific communication. However, currently there are hardly any systems which exploit the content of research articles for information retrieval. The paper describes our work carried out in developing ontology-based information retrieval system using information extracted about sentences in research articles. We present results of a supervised learning mechanism using conditional random fields for context identification and sentence classification of sentences in the related work section of research articles. The labelling of sentences is carried out based on a classification framework, which we propose for classifying sentences in these sections. We proceed to develop a sentence context ontology for modelling the classified data obtained through CRFs. We also show how the ontology is further used for creating RDF data. Finally, we describe the user interface developed using SEWESE tags and SPARQL for querying the developed RDF data.

EPrints makes its mark

Purpose – The purpose of the paper is to report on the impact and cost/benefit of implementing three EPrints digital repositories at the University of Otago, and to encourage others to follow suit.Design/methodology/approach – Three repositories were successfully implemented at the University of Otago using existing commodity hardware and free open source software. The first pilot repository was implemented within ten days, and is now a fully functional system that is being championed for institutional‐wide use by the University Library. The other two repositories emerged from different community needs. One is academic, concerned with collecting and researching indigenous content; the other is designed to preserve and manage collective memory and heritage content for a small rural community.Findings – The paper shows that digital repositories can be established quickly and effectively with surprisingly few resources; readily incorporate any kind of extant digital content, or non‐digital material that is c...

Contextual information retrieval in research articles: Semantic publishing tools for the research community

In recent years, the dramatic increase in academic research publications has gained significant research attention. Research has been carried out exploring novel ways of providing information services using this research content. However, the task of extracting meaningful information from research documents remains a challenge. This paper presents our research work on developing intelligent information systems that exploit online article databases. We present in this paper, a linked data application which uses a new semantic publishing model for providing value added information services for the research community. The paper presents a conceptual framework for modelling contexts associated with sentences in research articles and discusses the Sentence Context Ontology, which is used to convert the information extracted from research documents into machine-understandable data. The paper reports supervised learning experiments carried out using conditional probabilistic models for achieving automatic context identification. The paper also describes a Semantic Web Application that provides various citation context based information services.

Context identification of sentences in research articles: Towards developing intelligent tools for the research community

Scientific literature is an important medium for disseminating scientific knowledge. However, in recent times, a dramatic increase in research output has resulted in challenges for the research community. An increasing need is felt for tools that exploit the full content of an article and provide insightful services with value beyond quantitative measures such as impact factors and citation counts. However, the intricacies of language and thought, and the unstructured format of research articles present challenges in providing such services. The identification of sentence contexts that encode the role of specific sentences in advancing an article’s scientific argument can facilitate in developing intelligent tools for the research community. This paper describes our research work in this direction. First, we investigate the possibility of identifying contexts associated with sentences and propose a scheme of thirteen context type definitions for sentences, based on the generic rhetorical pattern found in scientific articles. We then present the results of our experiments using sequential classifiers – conditional random fields – for achieving automatic context identification. We also describe our Semantic Web application developed for providing citation context based information services for the research community. Finally, we present a comparison and analysis of our results with similar studies and explain the distinct features of our application.

Scalability of Techniques for Online Geographic Visualization of Web Site Hits

Extremely large data sets are now commonplace, and they are often visualized through the World Wide Web. Scalability of web-based visualization techniques is thus a key issue. This paper investigates the scalability of four representative techniques for dynamic map generation and display (e.g., for visualizing geographic sources of web site hits): generating a single composite map image, overlaying images on an underlying base map and two variants of overlaying HTML on a base map. These four techniques embody a mixture of different display technologies and distribution styles (three server-side and one distributed across both client and server). Each technique was applied to 20 synthetic data sets of increasing size, and the data set volume, elapsed time and memory consumption were measured. The results show that all four techniques are suitable for small data sets comprising a few thousand points, but that the two HTML techniques scale to larger data sets very poorly across all three variables.

Translating descriptions of a viewpoint among different representations

An important part of the systems development process is building models of real-world phenomena. These phenomena are described by many different kinds of information, and this diversity has resulted in a wide variety of modelling representations. Different types of information are better expressed by some representations than others, so it is sensible to use multiple representations to describe a phenomenon. This paper describes an approach to facilitating the use of multiple representations within a single viewpoint by translating descriptions of the viewpoint among different representations. An important issue with such translations is their quality, or how well they map the constructs of one representation to the constructs of another. Two methods are proposed for improving translation quality: heuristics and enrichment, and a preliminary metric for measuring relative translation quality is described.

Exploiting the advantages of object oriented programming in the implementation of a database design environment

We describe the implementation of a database design environment (Swift) that incorporates several novel features. Swifts data modelling approach is derived from viewpoint-oriented methods. Swift is implemented in Java, which allows us to easily construct a client-server based environment. The repository is implemented using PostgreSQL, which allows us to store the actual application code in the database. The combination of Java and PostgreSQL reduces the impedance mismatch between the application and the repository.

Precision of DXA-Derived Visceral Fat Measurements in a Large Sample of Adults of Varying Body Size: Precision of CoreScan and Association with BMI

The aim of this study was to determine the precision of GE Lunars CoreScan tool (GE Healthcare, Madison, Wisconsin) for measuring visceral adipose tissue (VAT) in adults of varying body size.

Characterising relational view updates using relative information capacity

In 2013, Chris Date published a book about the view update problem in the context of the Relational Model. He presented several detailed examples of different varieties of view updates and characterised their behaviour, in the process deriving a set of principles for updating views based on the notion of information equivalence of view schemas. In this paper we discuss work in progress that examines how Dates operational definition of information equivalence can be formally characterised using Hulls concept of relative information capacity. As a proof of concept, we use an extension of Millers schema intension graph formalism to model the information equivalence of Dates restriction view example.