Publication

Abstract : Stock price forecasting is a prominent topic in quantitative finance, as accurate prediction is essential due to the complexity of the market. This research work employs variational mode decomposition (VMD) to decompose stock data into several variational modes, further used to train a long short-term memory (LSTM) network with attention mechanism. The primary goal of this research is to enhance the accuracy of stock price prediction by exploring the effectiveness of VMD and attention mechanism techniques. From the experiment analysis, the efficacy of VMD is quantified as mean absolute error (MAE) score-163.91 and root mean square error (RMSE) score-192.39 from the results of LSTM with VMD. The efficacy of the attention mechanism is quantified as MAE score-94.16 and RMSE score-117.12 of from the results of VMD + LSTM + attention. The experimental results indicate that the application of VMD and attention mechanism to an LSTM model leads to improved predictions.