The open archive for STFC research publications
Full Record Details
Exploring the cloud of research information systematically
KG Jeffery (STFC)
A Asserson (UiB)
Imagine a cloud of data points (values) in multidimensional space. This is how research information appears commonly to end-users. There are many different kinds of data with varying quality, structured, semi-structured or unstructured with heterogeneous character sets, languages, syntax and semantics. Different end-users see the cloud differently: some data points are of more interest, some of less interest. The ?more-or-less? dimension has several aspects: the kind of data (entities), the values of data points (attribute-values) and the relationship between sets of data values (such as relationship role or relationship duration). As one example, a typical researcher is extremely interested in a scientific dataset and associated software together with the associated metadata describing the precision, accuracy, method of collection etc. The researcher is very interested in associated publications with interpretations of the scientific dataset. The researcher will be interested in who conducted the research, the rest of the team and the organisation where the research was done. The researcher has less interest in who funded the research, under which programme it was funded. From this we can construct a set of ?axes? or structural elements through the cloud along which data points of interest cluster. As another example, a research manager in a funding organisation may be interested in the value of funding awarded within one country to research in a particular subject area without being interested in to whom it was awarded nor the organisation where they work. This is a different view of the cloud with a different set of axes along which data points cluster. A research manager at a university might also be interested in output research publications by year, by department, by publication channel. As a final example consider the innovative entrepreneur. She will be interested in any products from or patents on research in a relevant topic across all countries, on the track-record of a researcher or her organisation in technology transfer and wealth creation and in any conditions attached to the research funding. This is yet another set of axes through the cloud. Three questions are paramount: (1) how to assure quality data (so that results are accurate); (2) how to assist the end-user in formulating correctly the query to obtain the expected results; (3) how to structure the data to obtain the optimal response in terms of performance, recall and relevance including taming heterogeneity. We demonstrate that all three are related. We outline a solution based on structured metadata, formal logic and knowledge engineering techniques exposed to the end-user as a user-friendly assistant with graphical metaphors.
Paper In Conference Proceedings
In Information Retrieval in a Current Research Information System (CRIS-IR), Copenhagen, Denmark, November 2006, (2006).
Exploring the Clo…ly Paper20060820.doc
Showing record 1 of 1
Login to Add and Manage Publications
Terms & conditions
Science and Technology Facilities Council
Switchboard: 01793 442000