| AKS: Ranking
of Relevant Authors
Based on the individual scores
of each document, AKS ranks the authors according to their relevance
in a specific document set, therefore it is easy to find experts
or potential referees in a certain field.
Example
Besides biomedical concetps,
people are also interesting. With this example we will see how we
can look to a certain field of knowledge, find the relevant authors
and check others fields where they have been doing research.
Fig 1. List of the most significant authors related to COX1 and COX2 ranked by
their relevance among the existing documents that relate to either of the
two enzymes |
COX-1 and COX-2 are two enzymes very important in pharmacologic
research. COX-1 is implicated in homeostatic functions (in stomach
and kidney) and COX-2 is implicated in inflammation processes.
There are several well-known COX inhibitor drugs like aspirin and
active on-going research to develop new COX-2-specific inhibitors
that do not interfere with COX-1 activity.
If we are interested in finding the most relevant authors contributing
to this area of research AKS can definitely help. AKS analyzes
the entire document corpus retrieved by a particular query and
extracts a list of authors ranked by relevance on a statistical
basis.
Fig. 1 displays a list of the most relevant authors of the documents
retrieved after performing a query to search for documents that
contain COX-1 or COX-2 (it is a conceptual search, i.e. the documents
can contain any of the synonyms for COX-1 and COX-2).
The most relevant author is "DuBois, R N".
What does he work on? In which aspects of COX enzymes has he
focused his research? What is the relevance of his work? AKS
helps to answer these questions as we can retrieve and analyze
the set of documents authored by this researcher.
The first step to answer this question is to collect all the
documents authored by DuBois, R N. We do a search and we meet the
first problem: the name of an author appears in several different
ways depending on the journal or even depending on the document
itself. We could find this author in several combinations according
to the presence or absence of the second name, the usage of initials
or the use of intermediate capitals in his last name.
As we are not sure of this author having
a unique way of being written, we search for the documents written
by any author beginning with "DuBois, R". As we can see in fig. 2, there are several "authors" matching
this "name". We reject those with the second initial or name not
beginning with N and check the remaining possibilities.
Fig 2. Ambiguity resolution for the authors���
names beginning with "DUBOIS, R". We select a set of different ways for
naming Raymond Dubois. Learn more about ambiguity resolution
at Powerful search
capabilities oriented to biomedical needs |
Fig 3. Result of the combined search of the different possible
ways Raymond Dubois is found in the scientific literature. |
The set of documents for this author
(whatever the "name" used
in the document) is displayed in fig. 3.
To take a closer look at DuBois work, we create an analysis of
the document set he has written or co-written (fig. 3, for more
information on document analysis see Large
document set analysis). From this study we can see the chemical
substances (fig. 4), genes (fig. 5) and diseases (fig. 6) that
are more significantly mentioned in his work. Beginning with the
diseases, things are very clear: his work is mainly centered in
colorectal cancer. Only marginally (item no. 13) appears "inflammation".
In the list of the genes (fig. 5), PTGS2 (also known as COX-2)
and PTGS1 (COX-1) are the most significant.
Fig 4. Cooccurrences yielded by the analysis (I): list of the
most relevant chemicals in the documents authored by Dubois,
R N. |
Fig 5. Cooccurrences yielded by the analysis (II): list of
the most relevant genes in the documents authored by Dubois,
R N. |
Fig 6. Cooccurrences yielded by the analysis (I): list of the
most relevant diseases in the documents authored by Dubois,
R N. |
|
The last point is to check the list of chemicals (fig. 4). The
three first are related to the chemical substances that are subtrates
(arachidonic acid, endoperoxid) and products (eicosanoids) of the
cyclooxygenases enzymes. Celecoxib and aspirin are COX inhibitor
drugs. In the case of celocoxib the inhibition is specific for
COX-2. They have been used to reduce the pain and inflammation
in general but what is their relation with colorectal cancer? This
question is answered if we check the documents containing these
chemicals.
We can focus our search towards the study of the NSAID (Non-steroidal
anti-inflammatory drugs) like aspirin or celecoxib in relation
with cancer. (for more information on focus searches see Suggestion
for focusing searches) We select those terms related with any
NSAID in the term list of the analysis result (fig. 7) and launch
a new analysis to rank the sentences taking into account the selected
terms. This analysis provides a set of meaningful sentences (fig.
8) that shows how DuBois defends the use of NSAIDs in colorectal
cancer prevention and cure.
Fig 7. Selection of the keywords (terms) related to NSAIDs
retrieved by the former analysis to make a second analysis
focused in these substances. |
Fig 8. Sentence list resulting from the second analysis (focused
in the NSAID). The sentences are ranked by their relevance
for the search and considering the term selection. Note
the high informative value of the sentences that are
listed. They summarizes the central points of the author's
work about the use of NSAIDs to inhibit COX-2 and their
role in the prevention and cure of the colorectal cancer. |
View
the TimeLogic
2006 Products,
including AKS (pdf)
Request more
information or schedule
a demonstration of AKS today!
|