关联模式识别与预测在证券市场的应用

The applicertion of the recognition and prediction of grey correlative pattern in the securities business

在吸收时间序列多重信息扩维原理和关联模式识别的思想的基础上,考虑到原关联度方法存在着诸多缺陷,故对其进行改造,提出了一种新的关联度方法:三角关联度方法.此方法的优点是将关联度的计算原则修改为先定性分析,后定量计算的原则,这样就克服了原关联度方法存在的种种缺陷,并将其应用于短期上证综合指数的预测,分析预测结果发现预测效果优良.

According to the grey correlative pattern recognition and the idea that an extended dimension approach from one dimension to multiple dimensions in time series,and considering the defects of the former computation models of the grey correlative degree,we reformed the former computation models and put forward a new computation model of the grey correlative degree： the triangle correlative degree.Boldly perfect the computational steps：qualitative analysis should be done before calculated.Consequently,the defects of the computation models of former correlative degrees had been overcome.Finally,applied the computation model to short forecasting in the securities business,analyzed the result and discovered that the effect of forecasting was very good.