Research on stock price prediction from a data fusion perspective

Abstract

<abstract> <p>Due to external factors such as political influences, specific events and sentiment information, stock prices exhibit randomness, high volatility and non-linear characteristics, making accurate predictions of future stock prices based solely on historical stock price data difficult. Consequently, data fusion methods have been increasingly applied to stock price prediction to extract comprehensive stock-related information by integrating multi-source heterogeneous stock data and fusing multiple decision results. Although data fusion plays a crucial role in stock price prediction, its application in this field lacks comprehensive and systematic summaries. Therefore, this paper explores the theoretical models used in each level of data fusion (data-level, feature-level and decision-level fusion) to review the development of stock price prediction from a data fusion perspective and provide an overall view. The research indicates that data fusion methods have been widely and effectively used in the field of stock price prediction. Additionally, future directions are proposed. For better performance of data fusion in the field of stock price prediction, future work can broaden the scope of stock-related data types used and explore new algorithms such as natural language processing (NLP) and generative adversarial networks (GAN) for text information processing.</p> </abstract>

Pricing and Optimization Model of European Options

“…This model has been improved in future uses and is a variant of the discrete Markov model. However, the traditional Hidden Markov Model has some room for improvement in the stock price prediction method [6]. In comparison to artificial intelligence techniques, HMM sets random initial parameters for customs clearance using the idea of data pattern recognition and iterates continually to find the optimal data model, which is more intelligent.…”