Title
Document Similarities in Web Based Collaborative Environments
Abstract
Collaborative editing of documents online in organizations helps the collaborators keep up to date with rapidly changing information, especially in technical fields. Because of its speed, editing documents on-line can have disadvantages, too: the repository of documents (document space) can grow large and redundant. To enforce coherence and reduce redundancy, a strong (hierarchical) structure of the document space can be imposed. In this case, each document has a strictly defined slot in the document space, it can be located by navigation and altered or extended. It has been our experience, however, that users do not like hierarchical structures and find the navigation cumbersome. Another possibility, more liked by the users, is to have a loosely structured document space, and support the users with tools for document retrieval and intelligent linking between documents. In this paper, we present a tool with which the user can see which other documents are related to his documents. The intended use for the tool is that the user edits a document to insert a query variable and after editing saves it. Saving a document automatically launches a query, whose results are a set of links to the most similar documents in the work space. In order to measure similarity, we have evaluated several methods, including Myers' O(ND) method (string comparson by edit distance) and a method based on Salton's Vector Model. In our reference document server (that contains about 1000 technical documents entered by the users) it was found that both methods worked quite well. The users of the document server have found our implementation useful, since they can (i) see if someone has already considered the topic that they are working with (ii) find information that is otherwise interesting to them.
Year
Venue
Field
2003
Frontiers in Artificial Intelligence and Applications
Discrete mathematics,World Wide Web,Web application,Mathematics
DocType
Volume
ISSN
Conference
105
0922-6389
Citations 
PageRank 
References 
0
0.34
1
Authors
3
Name
Order
Citations
PageRank
Michael Gindonis100.34
Tapio Niemi216318.90
Marko Niinimäki38311.43