A modified grey wolf optimization-based dendritic neural model for stock index return prediction

The prediction of stock index returns has always been a hot topic, which is an important basis for asset allocation and investment decisions of business managers and most investors. In order to improve the accuracy of the prediction, this paper proposes the use of a model based on the combination of the Improved Grey Wolf Optimization (IGWO) and Dendritic Neural Model (DNM). DNM has a more transparent structure, as well as a complex nonlinear processing capability and a unique architecture of automatic pruning, which can deal with the complexity of financial markets. For the traditional grey wolf optimization algorithm, this study designs four strategies to improve it. (i) Develop a nonlinear control parameter for dynamic equilibrium in global and local search algorithms. (ii) Chaos theory is introduced to plan the weights of wolves (iii) A local search scheme applicable to the head wolf is developed. (iv) The method of Levy leap is utilized to help the algorithm get rid of the local optimum. A total of 10 other algorithms are compared and validated based on data from 9 world representative stock index markets. It is pointed out that using IGWO in conjunction with DNM can achieve higher accuracy prediction.