Search

In recent years, the swift progress of artificial intelligence (AI) has significantly influenced trading practices, providing traders with advanced algorithms that improve decision-making and enhance trading strategies, leading to increased profits and reduced risks. The onset of the era of big data has further enriched this field, offering access to extensive financial data, such as historical stock prices, company financial statements, financial news articles, social media sentiments, and macroeconomic indicators—all publicly available. By identifying complex patterns and correlations within this vast data set, deep learning (DL) algorithms have proven their ability to predict stock prices and market trends more accurately than traditional methods. This comprehensive survey aims to provide an insightful examination of various deeplearning models employed in stock market forecasting. The primary objective is to categorize these models into two distinct types: Uni-modal and multimodal models. By exploring the nuances within each category, this literature survey provides a comprehensive understanding of these models’ strengths, applications, and contributions to the constantly evolving research landscape of stock market forecasting. Our survey adopts a systematic approach to categorize and analyze deep-learning models in stock market forecasting. Leveraging established databases and repositories, we will compile a comprehensive dataset comprising academic articles, conference papers, and other scholarly publications related to DL in finance. This dataset will span a defined period, allowing us to capture the temporal evolution of research trends in stock market prediction. The first phase involves extracting and compiling relevant literature from established databases, including but not limited to Scopus, Web of Science, and Google Scholar. This dataset will serve as the foundation for exploring the evolving landscape of DL applications in stock market forecasting. Subsequently, advanced techniques and methodologies will be employed to analyze citation patterns, model co-occurrence, and the intellectual structure of research in this domain. Our research identifies influential authors, collaboration networks, and geographical distribution of research activities to uncover emerging clusters of research excellence. The findings of this survey contribute valuable insights to both academia and industry. By categorizing and examining the strengths of uni-modal and multi-modal deep-learning models, researchers can refine their methodologies, and practitioners can make informed decisions regarding adopting predictive models in financial markets. Furthermore, the survey aims to guide future research directions, enhancing the overall effectiveness of predictive models in the dynamic landscape of stock market forecasting. In conclusion, this survey aims to provide a comprehensive overview of deeplearning models in stock market forecasting. By systematically categorizing and analyzing these models, our study aspires to contribute to the ongoing dialogue on integrating AI in financial practices, fostering a deeper understanding of the field’s evolution and future directions.

Assuming precast concrete using expansion material, we will confirm the influence of curing conditions on the basic physical properties of mortar with regard to the free expansion coefficient, constraint expansion coefficient and strength, using two types of expansive materials widely used in Japan . As a result, it was confirmed that the higher the curing temperature, the higher the strength regardless of the addition of the expansion material. In addition, it was confirmed that both the free expansion coefficient and the constrained expansion ratio had an effect irrespective of the curing temperature of the lime type expansion material.