基于GA-RS-LR算法的公司财务与个股投资探究

运用遗传算法-粗糙集-逻辑回归方法(GA-RS-LR)探讨我国A股上市公司财务与股票收益的关系.运用GA-RS方法获得财务指标最优约简;运用LR模型探求两者关系.最终,经GA-RS约简,60个财务指标中有17个对股票投资有重要影响;通过LR模型,4个指标具有显著效应;其中,负债与权益市价比为5.82%负效应,其余为正效应.对2015年股票相对波动进行预测得到70%的准确率,验证了GARS-LR模型对中长期投资的有效性.

Research on Corporate Finance and Stock Investment Based on GA-RS-LR Algorithm

To research the inner connection between the financial index system and the stock investment in 2014,data mining techniques of RS-Logit regression embedding genetic algorithm was applied.Through Rough Sets,the financial index system can be reduced by eliminating redundant variables.Local optima was successfully avoided by embedding Genetic Algo-rithm in RS reduction’s process.Furthermore,stepwise regression was applied to eliminate multi-collinearity of inputs,and Logit regression was established between the financial index system and the stock investment.The result shows that most inde-xes have positive effect on stock investment,except the ratio of debt and market price,with about 5.81 9% negative effect. 70% samples have a perfect forecast with data in 201 5,which has enhanced the effect of this model.