A multilabel dataset in machine learning comprises instances where each data point can possess multiple simultaneous labels or categories, capturing the intricate and overlapping nature of real-world classification scenarios. However, the high dimensionality inherent in multilabel datasets presents challenges, impacting aspects such as memory storage, computational costs, and model generalization. Despite ongoing efforts in developing MLFS approaches for classification, current methods struggle to address the diversity within feature sets, leading to a decline in model performance. To deal with this issue, our research introduces the first multimodal multiobjective evolutionary algorithm (MMDE_SICD) for MLFS in classification. By considering MLFS as a multimodal multiobjective problem, MMDE_SICD can effectively identify valuable feature subsets crucial for effective model training.

Multilabel learning is an emergent topic that addresses the challenge of associating multiple labels with a single instance simultaneously. Multilabel datasets often exhibit high dimensionality with noisy, irrelevant, and redundant features. In recent years, multilabel feature selection (MLFS) has gained prominence as a crucial and emerging machine learning task due to its ability to handle such data effectively. However, existing approaches for MLFS often prioritize top-ranked features based on intrinsic data criteria, disregarding relationships within the feature subset. Additionally, compared with conventional feature selection, multiobjective evolutionary algorithms (MOEAs) have not been widely explored in the context of MLFS. This study aims to address these gaps by proposing a multimodal multiobjective evolutionary algorithm (MMOEA) called MMDE_SICD which incorporates a preelimination scheme, an improved initialization scheme, an exploration scheme inspired by genetic operations ...