The Decision Model Of Stock Quantitative Investment Based On Rough Cognitive Networks

In recent years,with the development of computer application technology,quantitative investment has become one of the research hotspots in the financial field.Financial time series is a dynamic nonlinear complex system.How to analyze the changes in the forces of both the long and the short in the fluctuation of stock prices is still a challenging problem.In this paper,we propose a new approach to study this problem from the perspective of momentum investment,combined with Rough Cognitive Networks.The contributions and innovations of this article mainly include the following aspects:(1)For the analysis of financial time series,in this paper,a targeted optimization of the Rough Cognitive Networks model is proposed to the Extend Rough Cognitive Networks model.An evaluation method based on Rough Entropy is introduced into the Extended Rough Cognitive Networks,which can help the model to screen out the appropriate similarity threshold and effectively deal with the sensitive problem of similarity threshold.Meanwhile,an intensive analysis of the inference logic of the model is done to optimize the topology of the model.In addition,the three-way decision theory is combined with the reasoning ability of Fuzzy Cognitive Maps,and each decision is weighed during reasoning to arrive at the final preference recommendation.(2)Extended Rough Cognitive Networks are introduced into the momentum investment strategy of stock.First,the stock with a specific trend is screened out,and the relevant technical indicators are extracted.Secondly,the Rough Set theory is used for information granulation,and the reasoning structure of the model is constructed by combining the three-way decision with Fuzzy Cognitive Maps.Finally,the investment decision is made through the proposed decision method.(3)The attribute reduction method based on Neighborhood Rough Set is introduced into Rough Cognitive Networks,and the Neighborhood Rough Cognitive Networks(NRCNs)is proposed.Attribute reduction algorithm was used before the model reasoning to eliminate redundant attributes,reduce data dimensions,and improve computational efficiency.