Long Short-Term Memory Neural Network for Financial Time Series

Accounting journal entries as a long‐term multivariate time series: Forecasting wholesale warehouse output

“…The following papers presented in this section have been researched with the aim of collecting valuable knowledge about the nature of different kinds of data, the preprocessing phase, algorithms used in forecasting, their architecture, and finally, accuracy evaluation techniques used for performance measurement of different models. Some of the presented papers highlighted LSTM algorithms as the preferred choice of most researchers for financial time series forecasting Siami-Namini and Namin (2018) or Fjellström (2022), and other papers described text mining models that presented the best performance in comparison with statistical or machine learning models. Namely, Lashgar (2023) investigated the accuracy of text mining and technical analyses in forecasting financial time series (the S&P500 stock market index) and found that the FinBERT language model, designed for analyzing financial text data, outperformed the autoregressive integrated moving average (ARIMA) statistical model and the LSTM ML model, as evaluated by the root mean square error (RMSE) metric.…”

“…Sometimes, we may even need to use accounting knowledge and experience to further explore and refine metadata to ensure the accuracy of ML predictions for accounting forecasting targets. Despite the fact that literature quotes that recent developments in ML and NNs have given rise to nonlinear time series models that are increasingly being adapted for financial applications (Fjellström, 2022), the term financial data is widely used as in the described papers, and generally, it does not refer to accounting data from bookkeeping journals analyzed in this research. Accounting data are a broad term that includes various types of data ranging from…”

“…Despite the fact that literature quotes that recent developments in ML and NNs have given rise to nonlinear time series models that are increasingly being adapted for financial applications (Fjellström, 2022), the term financial data is widely used as in the described papers, and generally, it does not refer to accounting data from bookkeeping journals analyzed in this research. Accounting data are a broad term that includes various types of data ranging from stock prices and technical indicators to financial indicators from financial statements and data from accounting books.…”

Optimization of Traditional Stock Market Strategies Using the LSTM Hybrid Approach

“…It is particularly noteworthy that LSTM networks, compared to other models, provide impressive daily returns of 0.46% before transaction costs, thus confirming their efficiency and potential in financial forecasting. In the following paper [25], the authors present a ModAugNet model, which brings a new data augmentation approach for stock market index prediction using LSTM networks. Through testing on the S&P500 and KOSPI200 indices, ModAugNet showed a significant improvement in forecasting accuracy.…”

“…Central to this integration was the model's ability to recognize and interpret complex patterns in historical data, using them to create more accurate predictions. As a consequence of this integration, the trading strategy's ability to generate accurate trading signals should be improved, potentially contributing to better market performance when using machine learning methods [24,25]. Construction of the model and estimation process is highly important because it affects model efficiency [40].…”