Link Analysis in Mind Maps: A New Approach To Determine Document Relatedness
ABSTRACT Owner: Joeran, Added to JabRef: 2009.09.04
- SourceAvailable from: Joeran Beel
Conference Paper: Enhancing Information Search by Utilizing Mind MapsProceedings of the 21st ACM Conference on Hypertext and Hypermedia (HT'10); 01/2010
- [Show abstract] [Hide abstract]
ABSTRACT: An important aspect of a researcher's activities is to find relevant and related publications. The task of a recommender system for scientific publications is to provide a list of papers that match these criteria. Based on the collection of publications managed by Mendeley, four data sets have been assembled that reflect different aspects of relatedness. Each of these relatedness scenarios reflect a user's search strategy. These scenarios are public groups, venues, author publications and user libraries. The first three of these data sets are being made publicly available for other researchers to compare algorithms against. Three recommender systems have been implemented: a collaborative filtering system; a content-based filtering system; and a hybrid of these two systems. Results from testing demonstrate that collaborative filtering slightly outperforms the content-based approach, but fails in some scenarios. The hybrid system, that combines the two recommendation methods, provides the best performance, achieving a precision of up to 70%. This suggests that both techniques contribute complementary information in the context of recommending scientific literature and different approaches suite for different information needs.09/2014;
- [Show abstract] [Hide abstract]
ABSTRACT: Web search, academic search and expert search engines often have difficulties in classifying and ranking objects and generating summaries for them. In this paper I propose that data retrieved from mind maps may enhance these search applications. For instance, similar to anchor text analysis, documents, i.e. web pages or academic articles, linked in a mind map might be classified with text from the linking nodes. This idea, along with additional ideas, related work, the indented methodology, and first research results are summarized in this paper.
Mind Maps and Information Retrieval
University of Wolfenbüttel
Karl Scharfenberg Faculty
Abstract—Mind maps are used by millions of people. In this
paper we present how information retrieval on mind maps could
be used to enhance expert search, document summarization,
keyword based search engines, document recommender systems
and determining word relatedness. For instance, words in a mind
map could be used for creating a skill profile of the mind maps’
author and hence enhance expert search. This paper is a
research-in-progress paper which means no research results are
presented but only ideas.
Keywords-data mining, information retrieval, mind maps,
expert seach, document clustering, document classification
Mind maps were originally invented by Tony Buzan in the
1970s  and are nowadays used by millions of people for
brainstorming, note taking, project planning, decision making,
and document drafting. Many software tools exist to support
the creation of mind maps [2, 3]. The probably most popular
ones are MindManager with about 1.5 million users  and
FreeMind with about 150,000 downloads a month .
Hundreds of books and research articles were published about
how to create mind maps and about evaluating mind maps’
effectiveness, for instance, in the field of education [6-9].
MindMaps are also called Mind Maps Concept Maps or
However, to our knowledge, no research exists whether
information extracted from mind maps could be used for
enhancing other applications. We believe, it can and present
our ideas in this paper. Each of the next sections deals with
enhancing a particular application, namely
Keyword based search engines
(Document) Recommender Systems
Determining word relatedness
After presenting the ideas, the concept of information
retrieval on mind maps in general is discussed as well as future
Finding the right experts in a big company is a difficult
endeavor. In first attempts, databases were used and employees
could enter their skills manually [10, 11]. In the last decade
much research has been performed on automatically creating
skill profiles. The probably most promising approach is
analyzing documents. For instance, if a researcher has
published many documents containing the word ‘mind map’,
she probably has some expertise in the field of mind mapping.
Typical documents being analyzed are emails, visited websites,
scholarly articles and documents published in a company’s
intranet [12-16]. Mind maps have not been used so far.
Figure 1. A mind map (early draft of this paper)
A mind map (see Figure 1) seems well suited for creating a
skill profile of its author. The words in a mind map should
specify quite well the author’s expertise. In addition, nodes can
contain notes and links which could also be analyzed. In
contrast to text documents, a mind map seems likely to contain
less stop and other irrelevant words. This should facilitate the
creation of skill profiles.
III. DOCUMENT SUMMARIZATION
Search engines usually display summarized data for each
search result. This could be the document’s title, URL, or a
short extract of the document’s text. Academic search engines
additionally display data such as the author, publishing date or
the abstract (see Figure 2). This does not always deliver
satisfying results. In Figure 2, for instance, the extract is not
very informative, it equals basically the title. Alternatively to
text extracts, some researchers attempted to automatically
create abstracts [17-19] or summarizing documents based on
user generated data such as hyperlinks , social annotations
 and annotated bibliographies .
Figure 2. Example of summary data on Google Scholar
Mind maps could be used to complement summarization of
documents. Most mind mapping tools allow to link nodes in the
mind map with documents on the user’s hard drive or to link a
node to a webpage. The node’s text, and the text of parent
nodes, could be seen as a summary for the linked document.
Figure 3 illustrates this approach: The node with the red arrow
and gray background links to the PDF file of the scholarly
article ‘Are your citations clean?’ . This article deals with
problems of citation analysis. In this example the node which is
linking to the PDF and its parent nodes summarize the article’s
Citation Analysis -> Problems -> Technical Problems –>
problem: different authors with the same name
Certainly, one occurrence in a mind map would not be
sufficient for a thoroughly summary. But if several users would
link a document in several mind maps, this could add up to a
descent summary, highlighting what readers found most
relevant in the document.
Figure 3. Mind maps as document summary and for determining word
IV. KEYWORD BASED SEARCH ENGINES
When searching for documents, usually a keyword is
entered and the search engine returns those documents
containing the keyword. Various algorithms exist to calculate
how relevant a document is for a certain keyword search (e.g.
tf-idf and BM25(f)), but usually only words contained in the
documents are considered. Only few approaches consider
words of ‘neighbored’ documents
Considering neighbored documents means, document A could
be found for a keyword search even if document A does not
contain the keyword, but document B, which is linking to
document A. Usually this kind of link analysis is applied to
scholarly literature and websites. However, it seems likely that
the same concept could be applied to mind maps.
Figure 4. Enhance keyword based search engines
If a mind map links to a document, the words of the linking
(and parental) node could be assigned to the linked document.
Figure 4 illustrates this: The mind map contains a node called
‘expert search’ and child nodes link to documents related to
expert search (those with the red arrows). However, many of
these documents do not contain the term ‘expert search’, but
other expressions such as ‘expert finder’, ‘expertise
management’ or ‘skill management’. If search engines would
analyze mind maps and treat them as ‘neighbored’ documents,
recall in document retrieval could be increased.
(DOCUMENT) RECOMMENDER SYSTEMS
One common recommendation approach is to recommend
those items which are related to items a user likes (item based
recommendations). For scholarly literature and websites,
relatedness often is determined via citation analysis and
hyperlink analysis respectively. The same concept could be
applied to mind maps.
Node (No Link)
Node (No Link)
Node (No Link)
Node (No Link)
Node (No Link)
Figure 5. Expected Link Relatedness (Illustration)
The basic idea of what we call ‘Mind Map Citation
Analysis’: when two documents A and B are linked by a mind
map, document B could be recommended to those users liking
document A. This concept could be enhanced with common
citation analysis approaches. For instance, if two documents are
linked in high proximity, their relatedness can be expected to
be higher than two documents linked in lower proximity [25,
26]. Figure 5 illustrates this concept: Link 1 and 2 are in direct
proximity. Therefore, the linked documents can be expected to
be highly related. Between link 3 and 4 is a higher distance, so
their relatedness is likely to be lower.
VI. DETERMINING WORD RELATEDNESS
Knowing how words are related is important for many
applications. For instance, search engines want to determine
synonyms [27-29] and offer search query recommendations [-];
social tagging systems often recommend related tags to their
users [-]; and, among others, for web 2.0 applications, (semi)
automatic generation of ontologies is desirable [-]. Again, it
seems likely that information retrieval on mind maps could
help enhancing these applications.
A mind map is a graph and nodes are in hierarchical order.
As such the nodes’ terms are in direct relation to each other.
For instance, in Figure 3: Based on the mind map a search
query recommender could recommend the terms ‘problems’ or
‘definition’ to someone searching for ‘citation analysis’ in
order to specify his search. Or, if a person is searching for
‘citation analysis’, then ‘peer review’ might be an interesting
search term to broaden the search.
VII. DISCUSSION AND FURTHER RESEARCH
In this paper we presented how data of mind maps could be
used to enhance expert search, document summarization,
keyword based search engines, document recommender
systems and determining word relatedness. The presented ideas
are not yet supported by research and it could turn out that data
of mind maps is not able to enhance the mentioned
applications. In addition, two more challenges exist. First, it is
unknown if a sufficient number of people create mind maps
and if they are willing to share their data. Second, the
robustness of data seems critical. All platforms analyzing data
of users do have to cope with spam and fraud as soon as they
become successful. There is no reason to assume that this
would be different if information retrieval on mind maps
As part of the SciPlore.org project we will further research
information retrieval on mind maps. Recently we developed a
special mind mapping software focusing on researchers needs
. This software will help to gather and analyze mind maps in
order to see if the here presented ideas may be realized.
 Toni Buzan. Making the Most of your Mind. Pan Books, 1977.
 Open Directory Project DMOZ. Open Directory - Reference: Knowledge
Management: Knowledge Creation: Mind Mapping: Software. Website, Juli
 Mind-Mapping.org. Software for mindmapping and information
organisation. Website, Juli 2009. URL http://www.mind-mapping.org/mind-
 MindJet. MindJet: About MindJet. Website, Juli 2009. URL
 SourceForge. SourceForge.net: Project Statistics for FreeMind. Website,
 Sanford B. Frumker. Mind Map: Your Guide to Prosperity and
Fulfillment. Health Associates, 1993.
 Anthony J. Mento, Patrick Martinelli, and Raymond M. Jones. Mind
Mapping in Executive Education: Applications and Outcomes. The Journal of
Management Development, 4(4), 1999.
 Claire A. Michelini. Mind Map: A New Way to Teach Patients and Staff.
Home Healthcare Nurse, 18(5):318–322, 2000.
 Steve Morris and Jane Smith. Understanding mind maps in a week.
Hodder & Stoughton, 1998.
 T.H. Davenport and L. Prusak. Working knowledge: How organizations
manage what they know. Ubiquity, 1(24), 2000.
 ME Maron, S. Curry, and P. Thompson. An inductive search system:
Theory, design, and implementation. IEEE Transactions on Systems, Man and
Cybernetics, 16(1):21–28, 1986.
 C.S. Campbell, P.P. Maglio, A. Cozzi, and B. Dom. Expertise
identification using email communications. In Proceedings of the twelfth
international conference on Information and knowledge management, pages
528–531. ACM New York, NY, USA, 2003.
 N. Craswell, D. Hawking, A.M. Vercoustre, and P. Wilkins. P@noptic
expert: Searching for experts not just for documents. In Ausweb Poster
Proceedings, Queensland, Australia, 2001.
 B. Dom, I. Eiron, A. Cozzi, and Y. Zhang. Graph-based ranking
algorithms for e-mail expertise analysis. In Proceedings of the 8th ACM
SIGMOD workshop on Research issues in data mining and knowledge
discovery, pages 42–48. ACM New York, NY, USA, 2003.
 M. Maybury, R. D'Amore, and D. House. Expert finding for collaborative
virtual environments. 2001.
 J. Wang, Z. Chen, L. Tao, W.Y. Ma, and L. Wenyin. Ranking user's
relevance to a topic through link analysis on web logs. In Proceedings of the
4th international workshop on Web information and data management, pages
49–54. ACM New York, NY, USA, 2002.
 HP Luhn. The automatic creation of literature abstracts. Advances in
Automatic Text Summarization. MIT Press, Cambridge, MA, pages 58–63,
 CD Paice. The automatic generation of literature abstracts: an approach
based on the identification of self-indicating phrases. In Proceedings of the
3rd annual ACM conference on Research and development in information
retrieval, pages 172–191. Butterworth & Co. Kent, UK, UK, 1980.
 CD Paice. Constructing literature abstracts by computer: techniques and
prospects. Information Processing and Management: an International
Journal, 26(1):171–186, 1990.
 J.Y. Delort, B. Bouchon-Meunier, and M. Rifqi. Enhanced web document
summarization using hyperlinks. In Proceedings of the fourteenth ACM
conference on Hypertext and hypermedia, pages 208–215. ACM New York,
NY, USA, 2003.
 H. Zhang, Z.C.W. Ma, and Q. Cai. A study for documents summarization
based on personal annotation. In Proceedings of the HLT-NAACL 03 on Text
summarization workshop-Volume 5, pages 41–48. Association
Computational Linguistics Morristown, NJ, USA, 2003.
 M.Y. Kan, J.L. Klavans, and K.R. McKeown. Using the annotated
bibliography as a resource for indicative summarization. Arxiv preprint
 D. Lee, K. Jaewoo, M. Prasenjit, L. Giles, and O. Byung-Won. Are your
citations clean? Communications of the ACM, 50:33–38, 2007.
 X. Qi and B.D. Davison. Classifiers without borders: incorporating
fielded text from neighboring web pages. In Proceedings of the 31st annual
international ACM SIGIR conference on Research and development in
information retrieval, pages 643–650. ACM New York, NY, USA, 2008.
 Bela Gipp, Jöran Beel, and Christian Hentschel. Scienstein: A Research
Paper Recommender System. In International Conference on Emerging
Trends in Computing (ICETiC'09), pages 309–315. IEEE, January 2009.
Available on http://www.sciplore.org.
 Bela Gipp and Jöran Beel. Citation Proximity Analysis (CPA) - A new
approach for identifying related work based on Co-Citation Analysis. In
Proceedings of the 12th International Conference on Scientometrics and
Informetrics (ISSI'09), pages 571–575.
Scientometrics and Informetrics,
 V.D. Blondel, A. Gajardo, M. Heymans, P. Senellart, and P. Van Dooren.
A measure of similarity between graph vertices: Applications to synonym
extraction and web searching. Siam Review, 46(4):647–666, 2004.
 M. Clements, A.P. de Vries, and M.J.T. Reinders. Detecting synonyms in
social tagging systems to improve content retrieval. In Proceedings of the 31st
annual international ACM SIGIR conference on Research and development in
information retrieval, pages 739–740. ACM New York, NY, USA, 2008.
 Bela Gipp, Jöran Beel, and Ivo Rössling. ePassport: The World’s New
Electronic Passport. Createspace, Scotts Valley (USA), October 2007. ISBN
978-1434823182. Available on http://www.epassport-book.com/.