stock market index forecasting in iraq using rnn optimized by horse herd optimization algorithm

Stock market forecasting in emerging economies is a complicated task with large volatilities, reduced transparency, and the presence of various economic and political variables. the paper suggests an innovative hybrid model based on a recurrent neural network (rnn) integrated with the horse herd optimization algorithm (hhoa) for the directional movement forecasting of the iraqi stock market index. rnn extracts temporal dependencies from sequential financial information, and the hyperparameters of the network, i.e., learning rate and layer structure, are tuned by the hhoa for the purpose of optimization to gain better predictive accuracy. real-life datasets from the iraq stock exchange (isx) relating to the banking sector are utilized for the training and validation of the model. the suggested framework of the proposed rnn + hhoa is compared with the baselines of the kind like the rnn + particle swarm optimization (pso) and the rnn + genetic algorithm (ga). experimental results confirm that the proposed rnn + hhoa model outperforms the baselines consistently in accuracy, precision, recall, and f1-score. the paper suggests the scope for combining deep learning with the computational power of nature-inspired optimization algorithms for robust financial forecasting in highly volatile markets.