EPPI-Reviewer 4 Credits

Credits

EPPI-Reviewer is developed and maintained by James Thomas, Sergio Graziosi, Patrick O'Driscoll and Jeff Brunton.

It is built on Microsoft platforms, integrates various commercial and open source components and accesses third party web services. We wish to acknowledge and thank the following organisations and developers for their contribution to our effort and to the developers community in general.

Core Technologies:

EPPI-Reviewer 4 is written in Silverlight 5 using Microsoft Visual Studio (2015) and the Visual C# programming language.

EPPI-Reviewer Web is written in Angular using Microsoft Visual Studio (2017) and shares most of the backend C# code with version 4.

Data are stored on a Microsoft SQL Server (2017) Database.

Data exchange and business intelligence are handled via the Open Source (license) CSLA for Silverlight Framework created by Rockford Lhotka. Special thanks are due to Rockford for creating and maintaining this amazing piece of software.

User Interface: to enhance our user experience, several third party controls are integrated in EPPI-Reviewer user interface.

RadControls for Silverlight from Telerik provide the backbone for the EPPI-Reviewer 4 user interface.

EPPI-Reviewer Web uses numerous components, many of which come from the Telerik Kendo UI^® bundle.

EPPI-Reviewer Web includes a PDF viewer powered by PDFTron.

Diagrams are built using MindFusion DiagramLite.

Icons are taken form the Open Icon Library, a collection of over 10,000 Unique Icons. Free for anyone to use (licenses).

Microsoft Academic: We now link to a regularly updated copy of the Microsoft Academic dataset.

Records in this dataset can be searched and browsed in an interactive way, and also automatically identified to keep reviews constantly updated with new research as and when it is published. (See the developing literature on 'living' systematic reviews') We are currently developing and evaluating new machine learning tools to maximise the efficiency with which we locate relevant research in a rich dataset of more than 240 million records.

K. Wang et al., “A Review of Microsoft Academic Services for Science of Science Studies”, Frontiers in Big Data, 2019, doi: 10.3389/fdata.2019.00045

Term Extraction: the "Find similar documents" function relies on term extraction.

Terms can be extracted using external services:

Termine from the National Centre for Text Mining (NaCTeM) - University of Manchester.
Zemanta API.
Yahoo Query Language (YQL).

Behind the scenes: whenever possible, we've tried not to re-invent the wheel.

To save bandwith and speed up communications, all data exchange is compressed using the open source (license) SharpZipLib library.

The full-text search fuctions make some use of the code posted by E. W. Bachtal.

You can find a description of the code we have used at this link. Source code and (MIT-like) license are here.

Automatic document coding (clustering) is powered by Lingo3G by Carrot Search.

IN THIS SECTION