Understanding the Impact Of Epoch Levels On LSTM-Based Stock Price Forecasting
Sophia Wong
Gilbert Classical Academy, Gilbert, USA
Publication date: November 20, 2025
Gilbert Classical Academy, Gilbert, USA
Publication date: November 20, 2025
DOI: http://doi.org/10.34614/JIYRC2025II38
ABSTRACT
The aim of this study is to understand the relationship between epoch levels and the Long-Short Term Memory (LSTM) algorithm. This study will use LSTM to predict future stock prices of Microsoft, Apple, and Tesla where every test will have a different epoch value. The research results reveal that LSTMs performance heavily relies on and is influenced by epoch levels, but they do not showcase a clear direct or inverse relationship. A model’s predictions are heavily impacted by modifying epoch levels, demonstrating the riskiness and the precision required when identifying the correct epoch level. In order to obtain and ensure robust LSTM predictions, it is vital to find the most precise epoch level while avoiding overfitting or underfitting the LSTM predictions. Research results show that LSTM models require precise evaluation with its epoch coefficients in order to optimize the machine learning predictions of forecasting stock price movements.
The aim of this study is to understand the relationship between epoch levels and the Long-Short Term Memory (LSTM) algorithm. This study will use LSTM to predict future stock prices of Microsoft, Apple, and Tesla where every test will have a different epoch value. The research results reveal that LSTMs performance heavily relies on and is influenced by epoch levels, but they do not showcase a clear direct or inverse relationship. A model’s predictions are heavily impacted by modifying epoch levels, demonstrating the riskiness and the precision required when identifying the correct epoch level. In order to obtain and ensure robust LSTM predictions, it is vital to find the most precise epoch level while avoiding overfitting or underfitting the LSTM predictions. Research results show that LSTM models require precise evaluation with its epoch coefficients in order to optimize the machine learning predictions of forecasting stock price movements.
