自适应共振模型在股票分类中的应用

The Application of ART in the Classification of Security

采用针对聚类问题的神经网络方法,利用自适应共振(ART)模型,通过对我国证券市场中的股票进行实证分类,可以得出在给定数据解释能力方面,ART模型倾向于构建一个狭长的、具有某种函数关系的类;各类型股票分组的结果显示:收益率与标准差之间规律递增排列,市盈率效应和规模效应、主营业务增长率效应等一一显现。

Some given data explanation can be abstained by the means of the ART pattern by classifying the securities in our stock market. The ART pattern tends to make up a long and narrow classification of some functional relationship, a result show from the various kind of security, the increasing ranks between the earning rate and standard deviation, the profit efficiency of the market, the efficiency of scale and the increasing efficiency of the major business.