This is the Power point presentation for Semester 1 seminar which is the part of my 1st year M Tech course in Computer Science. It is completely based on the work done and the research paper by DOCEAR team. Thanks to them.

1. UTILIZING MIND-MAPS FOR INFORMATION
RETRIEVAL AND USER MODELLING
By Ms. Sunayana R. Gawde
M Tech in Computer Science
14109

2. ORIGINAL PAPER
 On Utilizing Mind-Maps for Information Retrieval
and User Modelling:
By:
 Joeran Beel
 Stefan Langer
 Marcel Genzmehr
 Bela Gipp

3. CONCEPT
 A mind map is a diagram used to visually organize
information. A mind map is often created around a
single concept and drawn as an image.
 Major ideas are connected directly to the central
concept, and other ideas branch out from those.
 As such they are often used for tasks including
brainstorming, project management and document
drafting.

4. EXAMPLE

5. TWO TYPES OF INFORMATION RETRIEVAL
APPLICATIONS, WHICH UTILIZED MIND-MAPS IN
PRACTICE.
 Search Engine for Mind Maps
By MindMeister and XMind
 User Modelling System-ads
By MindMeister and Mindomo

6. IDEAS FOR MIND-MAP BASED IR
APPLICATIONS

7. SEARCH ENGINES FOR MIND-MAPS
 Search Engines for Mind-Maps
 User Modelling
 Document Indexing / Anchor Text Analysis
 Document Relatedness
 Document Summarization
 Impact Analysis
 Trend Analysis
 Semantic Analysis

8. SEARCH ENGINES FOR MIND-MAPS:
 Mind-maps contain information that probably is not
only relevant for the given authors of a mind-map,
but also for others.
 Therefore a search engine for mind-maps might be
an interesting application.

9. USER MODELLING:
 Analogous to analyzing users’ authored research
papers, emails, etc., user modelling systems could
analyze mind-maps to identify users’ information
needs and expertise. User models could be used,
for instance, for personalized advertisements, or by
recommender systems, or expert search systems

10. DOCUMENT INDEXING / ANCHOR TEXT
ANALYSIS:
 Mind-maps could be seen as neighbouring
documents to those documents being linked in the
mind-maps, and anchor text analysis could be
applied to index the linked documents with the
terms occurring in the mind-maps. Such information
could be valuable, e.g., for classic search engines.

11. DOCUMENT RELATEDNESS:
 When mind-maps contain links to web pages or
other documents, these links could be used to
determine relatedness of the linked web pages or
documents. For instance, with citation proximity
analysis, documents would be assumed to be
related that are linked in close proximity, e.g. in the
same sentence. Such calculations could be
relevant for search engines and recommender
systems

12. DOCUMENT SUMMARIZATION:
 Mind-maps could be utilized to complement
document summarization. If a mind-map contains a
link to a web-page, the node’s text, and maybe the
text of parent nodes, could be interpreted as a
summary for the linked web page. Such summaries
could be displayed by search engines on their
result pages.

13. IMPACT ANALYSIS
Mind-maps could be utilized to analyze the impact
of the documents linked within the mind-map,
similar to PageRank or citation based similarity
metrics. This information could be used by search
engines to rank, e.g., web pages, or by institutions
to evaluate the impact of researchers and journals.

14. TREND ANALYSIS
 Trend analysis is important for marketing and
customer relation- ship management, but also in
other disciplines . Such analyses could be done
based on mind-maps. For instance, analyzing mind-
maps that stand for drafts of academic papers
would allow estimating citation counts for the
referenced papers. It would also predict in which
field new papers can be expected.

15. SEMANTIC ANALYSIS
 A mind-map is a tree and nodes are in hierarchical
order. As such, the nodes and their terms are in
direct relationship to each other. These
relationships could be used, for instance, by search
engines to identify synonyms, or by recommender
systems to recommend alternative search terms or
social tags.

16. FEASIBILITY

17. 1. NUMBER OF MIND-MAP USERS AND
(PUBLIC) MIND-MAPS

18. 2. CONTENT OF MIND-MAPS
 Analyzed the content of 19,379 mind-maps, created
by 11,179 MindMeister and Docear users.
 On average, mind-maps contained a few dozens of
nodes, each with two to three words on average.
 The number of links in mind-maps is low.
 Almost two thirds of the mind-maps did not contain
any links to files.

19. 3. USER ACCEPTANCE (EVALUATED WITH
SCIPLORE MINDMAPPING)

20. CONCLUSION

21. PROTOTYPE
 Click- through rate (CTR), i.e. the ratio of clicked
recommendations against the number of displayed
recommendations.
 Primarily used by researchers.
 Recommender system recommends research
papers
 Each time, a user modified, i.e. edited or created, a
node, the terms of that node were send as search
query to Google Scholar.

22. CTR BY NUMBER OF ANALYSED NODES

23. REFERENCES
 Beel, J., Langer, S., Genzmehr, M., Nürnberger, A.:
Introducing Docear’s Research Paper
Recommender System. Proceedings of the 13th
ACM/IEEE-CS Joint Conference on Digital Libraries
(JCDL’13). pp. 459–460. ACM (2013).
 Beel, J., Gipp, B., Langer, S., Genzmehr, M.:
Docear: An Academic Literature Suite for
Searching, Organizing and Creating Academic
Literature. Proceedings of the 11th International
ACM/IEEE conference on Digital libraries. pp. 465–
466. ACM (2011).

24. THANK YOU