Information Technology Reference
In-Depth Information
Case-based reasoning [11] is particularly applicable to this problem domain be-
cause it (i) supports a rich and evolvable representation of experiences, problems,
solutions and feedback; (ii) provides efficient and flexible ways to retrieve these ex-
periences; and (iii) applies analogical reasoning to solve new problems [27]. CBR
systems can be used to propose new solutions or evaluate solutions to avoid potential
problems. The chapter of CBR in health care is discussed in [13] [14], where the ad-
vantages of this paradigm are remarked. The research in [28] suggests that analogical
reasoning is particularly applicable to the biological domain, in part because biologi-
cal systems are often homologous (rooted in evolution). Moreover, biologists often
use a form of reasoning similar to CBR, where experiments are designed and per-
formed based on the similarity between features of a new system and those of known
systems. In [29] a mixture of experts for case-based reasoning (MOE4CBR) is pro-
posed. It is a method that combines an ensemble of CBR classifiers with spectral
clustering and logistic regression, but does not incorporates extraction of knowledge
techniques and does not focus on dimensionality reduction. Some approaches such as
[11] provide CBR solutions and knowledge extraction techniques, facilitating the
comprehension of the classification process. This chapter presents a CBR solution
which also incorporates new knowledge extraction techniques, but additionally fo-
cuses on the definition of innovative strategies for dimensionality reduction and clus-
tering. The following section presents a detailed account of the CBR system proposed
in this work.
3 CBR System as Paradigm for Classifying Microarray Data
This section presents the CBR system proposed in the context of this research and
provides a classification technique based on previous experiences for data from mi-
croarrays. The CBR developed system imitates the behaviour of human experts in the
laboratory and incorporates innovative knowledge discovery techniques. The system
receives data from the analysis of chips and is responsible for classifying individuals
based on evidence and existing data.
The purpose of CBR is to solve new problems by taking into account similar prob-
lems that were previously resolved in the past [10]. The primary concept when work-
ing with CBRs is the concept of case. A case can be defined as a past experience, and
is composed of three elements: a problem description which describes the initial prob-
lem; a solution which provides the sequence of actions carried out in order to solve
the problem; and the final stage which describes the state achieved once the solution
was applied. A CBR manages cases (past experiences) to solve new problems. The
way cases are managed is known as the CBR cycle, and consists of four sequential
steps which are recalled every time a problem needs to be solved: retrieve, reuse,
revise and retain. Each step of the CBR life cycle requires a model or method in order
to perform its mission.
In the CBR system proposed within this study, the retrieve phase filters variables,
and recovers important variables from the cases to determine the most influential for
the classification. Once the most important variables have been retrieved, the reuse
phase begins adapting the solutions for the retrieved cases to obtain the clustering.
Once this grouping is accomplished, the next step is knowledge extraction. The revise
Search WWH ::
Custom Search