Accurate Stock Price Forecasting Based on Deep Learning and Hierarchical Frequency Decomposition

Abstract

The stock market is playing an increasingly important role in the global economy. Accurate stock price forecasting not only aids the government in predicting economic trends but also helps investors anticipate higher expected returns. Nevertheless, hurdles such as nonlinearity, complexity and high volatility make it a daunting task to predict stock prices. To address this issue, this study proposes a new hybrid model, termed Hierarchical Decomposition-based Forecasting Model (HDFM), to decompose and forecast stock prices in a hierarchical fashion. The model utilizes complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) for the initial decomposition of stock price time series. To enhance the predictive efficiency, sub-series with similar sample entropy from the decomposition were combined using the K-means clustering method. Through a thorough analysis, it was found that the first combined sub-series contained more high-frequency signals. Therefore, the first combined sub-series is subjected to a second decomposition with variational mode decomposition (VMD). Subsequently, the gated recurrent unit (GRU) model was used to individually predict each sub-series. The final results were obtained by merging the prediction outcomes. The proposed model was evaluated on three different stock markets. The experimental results show that the proposed model outperforms other forecasting methods across all stock indices. Moreover, ablation studies demonstrated the effectiveness of each component within the proposed model.

Keywords

• Stock price prediction
• deep learning
• hierarchical decomposition
• clustering