Information Technology Reference
In-Depth Information
occurring. Several systems under development aim at being placed in routine
clinical use. One of the most interesting impacts of CBR on the health sciences
lies in the position CBR must find with respect to statistics, which is routinely
used for data analysis and processing in experimental sciences. This is a major
trend in CBR in the Health Sciences research, as described next.
5.3 CBR versus Statistics in the Health Sciences
In health sciences domains, statistics is considered to be the scientific method
of choice for data analysis. Therefore, CBR in the Health Sciences systems re-
searchers have studied how to position CBR in these domains in relation to
statistics. Biometry is “the application of statistical methods to the solution of
biological problems” [59]. Statistics itself has several meanings. A classical defini-
tion of statistics is “the scientific study of data describing natural variation” [59].
Statistics is generally used to study populations or groups of individuals: it deals
with collections of data, not with single data points. Thus, the measurement of
a single animal or the response from a single biochemical test will generally not
be of interest; unless a sample of animals is measured or several such tests are
performed, statistics ordinarily can play no role [59]. Another main feature of
statistics is that the data are generally numeric or quantifiable in some way.
Statistics also refers to any computed or estimated statistical quantities such as
the mean, mode, or standard deviation [59].
The origin of statistics can be traced back to the seventeenth century, and
derives from two sources. One is related to political science and was created to
quantitatively describe the various affairs of a government or state. This is the
origin of the term “statistics.” In order to deal with taxes and insurance data,
problems of censuses, longevity, and mortality were studied in a quantitative
manner [59]. The second source is probability theory, which was also developed
in the seventeenth century, spurred by the popular interest in games of chance
among upper society (Pascal, de Fermat, Bernouilli, de Moivre) [59]. The sci-
ence of astronomy also fostered the development of statistics as a mathematical
tool to build a coherent theory from individual observations (Laplace, Gauss).
Applications of statistics to the life sciences emerged in the nineteenth century,
when the concept of the “average man” was developed (Quetelet) along with the
concepts of statistical distribution and variation [59]. Statistics researchers focus
on summarizing data: “All these facts have been processed by that remarkable
computer, the human brain, which furnishes an abstract” [59]. Statistics in-
volves reducing and synthesizing data into figures representing trends or central
tendencies.
There are actually two approaches in statistics, experimental and descriptive.
The experimental approach, at the basis of any theory formation in experimental
sciences, and in particular the life sciences, refers to a method aimed at identify-
ing relations of cause to effect. A statistical experiment needs to follow a precise
and controlled plan with the goal of observing the effect of the variation of one or
more variables on the phenomenon under study, while eliminating any potential
hidden effects. The statistician is responsible for the design of the experiment
Search WWH ::
Custom Search