Adaptive Multi-Tree Genetic Programming for Multiple Feature Construction

Created by W.Langdon from gp-bibliography.bib Revision:1.8592

@InProceedings{niu:2025:GECCOcomp,

• author = "Jiaxin Niu and Xiaoying Gao and Minghui Bai and Jianbin Ma",
• title = "Adaptive Multi-Tree Genetic Programming for Multiple Feature Construction",
• booktitle = "Proceedings of the 2025 Genetic and Evolutionary Computation Conference Companion",
• year = "2025",
• editor = "Aniko Ekart and Nelishia Pillay",
• pages = "651--654",
• address = "Malaga, Spain",
• series = "GECCO '25 Companion",
• month = "14-18 " # jul,
• organisation = "SIGEVO",
• publisher = "Association for Computing Machinery",
• publisher_address = "New York, NY, USA",
• keywords = "genetic algorithms, genetic programming, multi-tree, adaptive, feature construction, multiple feature: Poster",
• isbn13 = "979-8-4007-1464-1",
• URL = "https://doi.org/10.1145/3712255.3726595",
• DOI = "doi:10.1145/3712255.3726595",
• size = "4 pages",
• abstract = "Feature construction has been proven to be an effective method for improving classification performance. Genetic Programming (GP) has demonstrated its immense potential in automatic feature construction due to its flexible representation. Multi-tree GP can enhance the search ability and efficiency of GP, and has been used for feature construction. However, existing multi-tree methods often require a predetermined number of trees in each individual, which limits the flexibility and adaptability of the algorithm. In this paper, a novel Adaptive Multi-tree GP (AMTGP) based feature construction method is proposed, which can automatically adjust the number of trees in each individual based on the current state and fitness feedback. Experiments on 10 datasets show that our proposed AMTGP can achieve better classification performance than the single-feature construction method (SFC). Compared to the traditional Multi-tree GP (MTGP) methods, AMTGP not only reduces the workload of parameter tuning but also enhances the algorithm's adaptability to different datasets, resulting in better classification performance. Compared to four state-of-the-art benchmark techniques, in most cases, our proposed method can achieve better classification performance.",
• notes = "GECCO-2025 GP A Recombination of the 34th International Conference on Genetic Algorithms (ICGA) and the 30th Annual Genetic Programming Conference (GP)",

}

Genetic Programming entries for Jiaxin Niu Xiaoying (Sharon) Gao Minghui Bai Jianbin Ma

Citations