Data-Driven Precision by a Novel Fusion of Deer Hunting Optimization and LSTM Networks for Stock Price Forecasting

This study introduces a novel hybrid approach of utilizing Long Short-Term Memory Neural Networks and a completely new optimization technique called Deer Hunting Optimization. This model referred to as the Deer Hunting Optimization-Long Short-Term Memory is an enhancement to the long short term memory network since it improves prediction accuracy through the optimization of the parameters of the Long Short Term Memory Network. The study used data from the Dow Jones Industrial Average and the various models were compared to other configurations such models optimized with Particle Swarm Optimization or Genetic Algorithm techniques and beyond, three other artificial neural network models. The Deer Hunting Optimization-Long Short-Term Memory model obtained results of a mean squared error of 7.19, a mean absolute error of 2.7 and an R-square value of 0.44 respectively which is better than other models. According to these results, it is shown that the DHO-LSTM model can predict stock prices successfully and tuning the parameters with Deer Hunting Optimization achieves 15–20% improvement in prediction accuracy. Due to this the scope of how AI along with optimization can be used to better the prediction of stock performance, and its applications, like active trading, looks good and can be transferable to other markets with time constraints.