Superior
prediction and classification in determining a firm's
performance are the major concerns for practitioners
and academicians as it is important to provide useful
information to shareholders and potential investors
to enable them make good decisions regarding investments.
A firm's performance can be analyzed based on financial
indicators reported in that company's annual report;
balance sheet, income, and cash flow statements. These
reports provide vast amount of information related to
the performance of the firm. To provide a better understanding
of the performance of a particular firm, these financial
data are best transformed into financial ratios. Nevertheless,
there are a number of financial ratios to be considered
in classifying the performance of each firm.
Some of
these financial ratios may be irrelevant or correlate
with one another giving redundant information for classification.
In addition, different studies or different firms use
different financial ratios in analyzing firms' performance.
The level of importance of the ratios differs from one
research to another, industry to industry, and from
one country to another. Using all the ratios to build
the classification model will cause a system to operate
in high dimension. This will complicate the learning
process computationally and analytically (Pechenizkiy
et al., 2006). In addition, regression method
and neural networks are difficult to use when the number
of features are large (Lendasse et al., 2001).
Thus, it is crucial to discover the significance of
the financial ratios because of its impact on the accuracy
and classification of the model developed. |