Studies on stepwise discriminant analysis methodology based on statistic data

Discriminant Analysis is a kind of statistic arithmetic to predict the group membership of new observations (samples) based on a linear function which was constructed by observations whose group membership had already been known. To identify variables effectively, stepwise discriminant analysis arithmetic was proposed in this paper. By added or removed variables, stepwise discriminant analysis method can filter out unimportant variables when identify observations´ group membership. A case analysis about regional economy development discriminant issue was also proposed to testify the discriminant capability of the stepwise discriminant analysis arithmetic.