Bird's Eye View feature selection for high-dimensional data.

Area of Science:

• Machine Learning
• Data Science
• Computational Intelligence

Background:

• High-dimensional datasets in machine learning often contain irrelevant features, noise, and outliers.
• These data issues degrade model performance and increase computational costs.
• Effective feature selection is vital for building robust and efficient machine learning models.

Purpose of the Study:

• To introduce and evaluate the Bird's Eye View (BEV) feature selection technique.
• To address the challenges posed by high-dimensional data in machine learning.
• To improve classification performance while reducing the number of selected features.

Main Methods:

• The Bird's Eye View (BEV) technique integrates Evolutionary Algorithms (Genetic Algorithm), Dynamic Markov Chain, and Reinforcement Learning.
• A population of agents is maintained and guided through the feature search space.
• Agents are rewarded or penalized based on their performance in feature selection.

Main Results:

• The BEV technique demonstrated superior classification performance compared to conventional methods.
• BEV significantly reduced the number of features required for accurate predictions.
• The proposed method outperformed existing state-of-the-art feature selection techniques on benchmark datasets.

Conclusions:

• The Bird's Eye View (BEV) is an effective feature selection strategy for high-dimensional data.
• BEV offers a promising approach to enhance machine learning model efficiency and accuracy.
• This novel technique provides a biologically inspired solution to a core machine learning challenge.