Quadratic descriptors and reduction methods in a two-layered model for compound inference

• DOI: 10.3389/fgene.2024.1483490
• Publication Date: 2025-01-29T07:01:25Z
• Authors: Jianshen Zhu; Naveed Ahmed Azam; Shengjuan Cao; Ryota Ido; Kazuya Haraguchi; Liang Zhao; Hiroshi Nagamochi; Tatsuya Akutsu
• Keywords: Overfitting

Compound inference models are crucial for discovering novel drugs in bioinformatics and chemo-informatics. These rely heavily on useful descriptors of chemical compounds that effectively capture important information about the underlying constructing accurate prediction functions. In this article, we introduce quadratic descriptors, products two graph-theoretic to enhance learning performance a two-layered compound model. A mixed-integer linear programming formulation is designed approximate these inferring desired with Furthermore, different methods reduce aiming avoid computational complexity overfitting issues during process caused by large number descriptors. Experimental results show 32 properties monomers 10 polymers, functions constructed proposed method achieved high test coefficients determination. our inferred time ranging from few seconds approximately 60 s. indicate strong correlation between graphs their