Automatically Choosing Selection Operator Based on Semantic Information in Evolutionary Feature Construction

Hengzhe Zhang, Qi Chen, Bing Xue, Wolfgang Banzhaf, Mengjie Zhang

Abstract: In recent years, genetic programming-based evolutionary feature construction has shown great potential in various applications. However, a critical challenge in applying this technique is the need to select an appropriate selection operator with great care. To tackle this issue, this paper introduces a novel approach that leverages the Thompson sampling technique to automatically choose the optimal selection operator based on semantic information of genetic programming models gathered during the evolutionary process. The experimental results on a standard symbolic regression benchmark containing 37 datasets show that the proposed adaptive operator selection algorithm outperforms expert-designed operators, demonstrating the effectiveness of the adaptive operator selection algorithm.