BGOA-TVG: Binary Grasshopper Optimization Algorithm with Time-Varying Gaussian Transfer Functions for Feature...

Area of Science:

• Machine Learning
• Data Mining
• Swarm Intelligence

Background:

• Feature selection is crucial for improving classification accuracy in machine learning and data mining.
• Traditional binary optimization algorithms often use S-shaped or V-shaped transfer functions, which may limit convergence speed and global search capability.

Purpose of the Study:

• To propose a novel Binary Grasshopper Optimization Algorithm using time-varying Gaussian transfer functions (BGOA-TVG) for effective feature selection.
• To evaluate the performance of BGOA-TVG against traditional and state-of-the-art swarm intelligence algorithms.

Main Methods:

• Development of the BGOA-TVG algorithm incorporating time-varying Gaussian transfer functions to map continuous search spaces to binary ones.
• Comparative analysis of BGOA-TVG with S-shaped and V-shaped binary grasshopper optimization algorithms and five other swarm intelligence algorithms.
• Testing the algorithms on benchmark datasets: UCI, DEAP, and EPILEPSY.

Main Results:

• The proposed BGOA-TVG demonstrated a faster convergence speed and stronger global search capability compared to traditional transfer functions.
• BGOA-TVG achieved superior performance in feature selection across the UCI, DEAP, and EPILEPSY datasets.
• Experimental results indicate BGOA-TVG's effectiveness in identifying crucial features for enhanced classification accuracy.

Conclusions:

• The BGOA-TVG algorithm offers an effective and efficient approach for feature selection in machine learning.
• Time-varying Gaussian transfer functions provide significant advantages in optimizing binary search spaces for swarm intelligence algorithms.
• BGOA-TVG represents a promising advancement for improving classification accuracy through optimized feature selection.