Transactions on Case-Based-Reasoning

This journal is the second journal in the se ries of journals Transaction on Case-Based
Reasoning. It is comprised of papers presented at the second workshop on Case-Based
Reasoning on Multimedia Data CBR-MD 2009 (www. cbr-md. org) and selected topics
on similarity from the Industrial Conference on Data Mining ICDM 2009
(www. data-mining-forum. de).
It starts with a paper on distances in classification by Weihs and Szepannek that
gives a fast introduction on the problem of similarity respective distances and how to
find out the best measure for the current problem. Classification is seen as supervised
and unsupervised classifica tion so that it can be viewed from the clustering and
classification aspect. They explain on different applications how to find out the best
similarity.
Their paper brings the methodology how to deal with similarity for different
applications closer to the engineers and praticioners and will help to pave the way for
more Case-Based Reasoning applications.
The next paper by Vorobieva and Schmidt deals with the problem of exceptional
cases that are handled by Case-Based Reasoning and the usual missing data problem.
This is a big problem in particular in medical applications. The approach they
propose combines Case-Based Reasoning with a statistical model. The statistical
model is used to summarize the know cases while Case-Based Reasoning is used to
explain the exceptional cases.
A system that uses Case-Based Reasoning to diagnose Lymphatic Tumors on
Microscopic Images is presented by Colantonio et. al. The analysis of cells in
microscopic images for different medical diagnosis purposes is a hot topic in
medicine. Recently a lot of alternative methods, very specific and expensive chemical
or molecular biological tests, are proposed to come around visual microscopic image
diagnosis to avoid the problem that goes along with visual inspection. Altought most
visual microscopic tests are good for mass screening, automation is still limited.
There are not enough automatic image interpretation system available that can help to
automatize these problems. A case-based reasoning system with proper functions can
quickly adapt an image interpretation system to different problems.
Meta Learning of System Parameters by Case-Based Reasoning is proposed by
Attig and Perner. It is a big problem in signal and image analysis is to find a model
for the signal transformation that holds not only for a few images rather to a class of
images. Case-Based Reasoning with its incremental learning functions can help to
solve this problem.