Children’s
Information Retrieval
The
digital world is the go-to place for obtaining information of any
kind. Even kids prefer using digital libraries over physical
libraries now. The classroom is no longer the primary source of
knowledge for the current generation of kids.
The
interfaces of these digital libraries are designed by the adults
based on their personal search experiences. However, the search
preferences for a kid vary markedly in comparison to those of an
adult, making the results of search engines designed for adults
mostly irrelevant for children.
Isn’t
it therefore of paramount importance that search interfaces be
designed for children in accordance with their personal needs so that
they can get the most relevant information?
The
term “Children’s Information retrieval” points to the issue of
getting the most relevant information for children based on their
search queries.
Most
children grow up using computers, and are often even better equipped
than many adults to handle computer-based work. However, researchers
have discovered many differences between the searching strategies of
adults and those of children. Some of the differences are:
-
Children generally make more spelling mistakes.
-
Children scroll less than adults.
-
Adults have a more systematic navigation style.
-
Children use different relevance criteria from those used by adults.
-
Children have different concepts and categories in mind.
-
Children write shorter and more concise queries.
On
the basis of the above mentioned differences, any search query can be
recognised as being performed by an adult or by a child. An analysis
of the differences also helps develop a better understanding of the
information needs of children.
Another
key issue in Children’s Information Retrieval is query formulation.
Query formulation strategies depend on the thinking process of
children. To further complicate the issue, even these strategies vary
significantly from one age group to another. It is probably not
appropriate to categorise all children (people under the age of 16)
into one category, since the brain’s ability to comprehend
information changes drastically in this growing period. To better
understand the thinking process of children, surveys, interviews,
task analysis, focussed group discussions and other similar
activities should be performed. These activities will be helpful in
drawing inferences about the thinking process of children belonging
to several different age groups. Child psychologists can also help us
better understand how the brain of a child works.
The
last core issue in Children’s Information Retrieval is “ranking”.
How should we rank the results so that the documents most relevant to
children appear in the top K results? This is probably the most
important question, as the performance of the system is dependent on
the precision and efficiency of the topmost results.
Let’s say, we have two documents which both contain articles written on a particular topic. Doc1 consists of vocabulary terms frequently used in daily life, while Doc2 consists of many technical and uncommon terms. Which document should be preferred if the search has been performed by a child? The answer is quite intuitive. A child will definitely prefer Doc1, since it will be easier for him/her to comprehend. In most IR systems, documents with high idf scores (more rare terms) are preferred over documents that have a collection of commonly used terms. However, basic intuition suggests that for a child, simple vocabulary wins out over complex vocabulary, and preference should be given to documents which consist of a lesser number of rare terms. A high tf score and a low idf score should help a document obtain a higher rank.
Let’s say, we have two documents which both contain articles written on a particular topic. Doc1 consists of vocabulary terms frequently used in daily life, while Doc2 consists of many technical and uncommon terms. Which document should be preferred if the search has been performed by a child? The answer is quite intuitive. A child will definitely prefer Doc1, since it will be easier for him/her to comprehend. In most IR systems, documents with high idf scores (more rare terms) are preferred over documents that have a collection of commonly used terms. However, basic intuition suggests that for a child, simple vocabulary wins out over complex vocabulary, and preference should be given to documents which consist of a lesser number of rare terms. A high tf score and a low idf score should help a document obtain a higher rank.
This
topic is still in the early stages of research. Information retrieval
systems should be developed considering the needs of children, as
more and more children are using digital libraries to expand their
knowledge. The digital world should become more accessible for the
future generation of thinkers. Let us nurture the Einsteins of this
world to make this world a better place.
References
[1]
Hanna
Jochmann-Mannak, Theo Huibers, and Ted Sanders. 2008. Children's
information retrieval: beyond examining search strategies and
interfaces ”In Proceedings
of the 2nd BCS IRSG conference on Future Directions in Information
Access (FDIA'08)”
BCS Learning & Development Ltd., Swindon, UK, 8-8.
Comments
Post a Comment