Effective forecasting of stock market price by using extreme learning machine optimized by PSO-based group oriented...

Area of Science:

• Computational Finance
• Machine Learning
• Financial Forecasting

Background:

• Accurate stock index price forecasting is crucial for investors and financial analysts to optimize decision-making for maximum profit and minimal risk.
• Existing forecasting models often lack the robustness required for reliable financial predictions, necessitating advanced computational approaches.

Purpose of the Study:

• To enhance stock market forecasting effectiveness by integrating a modified Crow Search Algorithm (CSA) with Extreme Learning Machine (ELM).
• To introduce and evaluate a novel hybrid model, the Particle Swarm Optimization (PSO)-based Group oriented CSA (PGCSA) combined with ELM (PGCSA-ELM), for predicting stock index prices.

Main Methods:

• Development of a modified CSA algorithm, termed PGCSA, incorporating Particle Swarm Optimization (PSO) principles for enhanced performance.
• Application of the PGCSA algorithm to optimize the weights and biases of the Extreme Learning Machine (ELM) model.
• Validation of the hybrid PGCSA-ELM model using performance metrics, technical indicators, and hypothesis testing (paired t-test) on seven stock indices, including data from the COVID-19 period.

Main Results:

• The PGCSA algorithm demonstrated superior performance compared to existing algorithms on 12 benchmark problems.
• The hybrid PGCSA-ELM model showed significant effectiveness in predicting the next day's closing prices for seven diverse stock indices.
• Comparative analysis against existing techniques confirmed the robustness and accuracy of the proposed PGCSA-ELM model.

Conclusions:

• The PGCSA-ELM model represents a suitable and effective tool for accurate stock index price forecasting.
• The integration of PSO-enhanced CSA with ELM offers a promising approach for improving financial market prediction capabilities.
• The model's performance during the COVID-19 outbreak highlights its potential utility in volatile market conditions.