|
Improvement in computer technology have helped to develop
statistical applications with significant promotion and
evolve. Collect and record data is a needful work for
statistical analysis. However, the data files which recorded
integrity data are more and more huge by the development of
human civilization. But we usually only need parts of the
data, namely the subsets of the original data files when we
analyzing these data. So, how to take the subset efficiently
is very important.
To improve the taking of data, the most efficiently way is not
to ameliorate the method of getting it, but is to ameliorate
the method of saving it. Therefore,have an idea in my mind that
is "managing data dispersed"─To disperse big data file to
several small ones, and use a file to record the indexes of
these small data file. By the fitting connect, we can search
and analyze data quickly. This is a more quick and more
efficient way than searching from a huge data file, and it can
prevent the difficult in safeguard data because of the data file
is too large.
However, if we want to reach this goal, we should depend on the
statistical packages. But user must know how to operate and
familiar with the statistical packages, and can order the
instructions or design programs. And if we want to get the
results which are more fitting what we need, it must depended on
more complicated instructions or programs. Unless the
manager is very familiar with the statistical package,
otherwise, he may keep away from it.
|