Cuckoo search optimized integrated framework based on feature clustering and deep learning for daily stock price forecasting

Abstract

Under the background of the increasing prosperity of Internet finance, quantitative investment has become a hot topic, among which the prediction of stock price is the focus of research. In this paper, an optimized nonlinear integration framework based on feature clustering and deep learning is proposed to predict stock price daily data. Clustering algorithm is used to divide the complex and changeable stock price data into multiple clusters according to its characteristics, which can pave the way for the establishment of forecast model. Bidirectional long short-term memory (BiLSTM) network is introduced to construct the core of the proposed framework for accurately extracting timing information. Finally, the Radial basis function neural network based on Cuckoo search optimization (CSO-RBF) is constructed for final integration, which shows obvious advantages in improving generalization ability and adaptability of the model. The simulation results, compared with other benchmark models, demonstrate that the prediction performance of the proposed optimization integration framework is obviously better, which provides an effective method for stock price prediction.