USING LSTM DEEP LEARNING MODEL IN STOCK PRICE PREDICTION

Abstract

In recent times, the stock market has attracted a diverse range of participants, from organizations to financial experts, each employing different investment strategies. The common objective of investors is to maximize returns through investment. While various prediction methods have been proposed to mitigate risks, the application of artificial intelligence in stock price prediction continues to garner attention and research interest. Particularly, predicting time series data with irregular, non-seasonal characteristics, such as stock price data, remains a challenging task. This paper presents a method utilizing the Long Short-Term Memory deep learning model for stock price prediction and provides a comprehensive review of this model. The results indicate that the proposed method can predict stock price trends of adjusted closing prices with a root mean square error  of 0.1387 and mean absolute error of 0.1007. Although the Long Short-Term Memory method may not achieve highly accurate predictions, it can offer a reasonably close approximation to real-world data trends.