Enhancing Stock Market Prediction Accuracy Through Advanced Metaheuristic Techniques and Macro-Economic Factors Integration

This research purposes to explore and develop advanced metaheuristic methods to optimize machine learning unambiguously in ''Macro-Economic Factors'' (MEFs) price prediction. The principal goal is to increase the precision and efficacy of machine learning models employed in forecasting MEFs prices. This attempt involves discovering, designing, and implementing advanced metaheuristic algorithms proficient of successfully optimizing machine learning model constraints. The study emphases on enhancing ''Stock Market'' (SM) forecast precision through five innovative methods: an upgraded metaheuristic algorithm, metaheuristic algorithm-optimized ''Extreme Learning Machine'' (ELM), ''Intrinsic Time-scale Decomposition'' (ITD) based optimized ELM, ''Fuzzy Inference System based ELM'' (ELMFIS), and the addition of MEFs. Anticipated techniques such as ''Sine Cosine Algorithm optimized Fuzzy Symbiotic Organism Search'' (SCAOFSOS) and ''Particle Swarm Optimization-based Group-oriented Crow Search Algorithm'' (PGCSA) are showcased as effective tools, with reasonable analysis supporting PGCSA for its preeminence. These metaheuristic practices are further applied to assembly ELM, thereby attractive forecast accurateness. The amalgamation of ITD with ''Cluster-based Modified Crow Search Algorithm'' (CMCSA) optimized ELM validates superior predictive abilities. The study's result highlights the substantial impact of MEFs, such as money exchange rates and crude oil prices. As a result, SM dynamics, with SM price forecast notably enhanced by integrating real and forecast MEFs along with conventional stock-index parameters.