Modified marine predators algorithm for feature selection: case study metabolomics

Feature selection (FS) is a necessary process applied to reduce the high dimensionality of the dataset. It is utilized to obtain the most relevant information and reduce the computational efforts of the classification process. Recently, metaheuristics methods have been widely employed for various optimization problems, including FS. In the current study, we present an FS method based on a new modified version of the marine predators algorithm (MPA). In the developed MPASCA model, the sine-cosine algorithm (SCA) is utilized to improve the search ability, which works as a local search of the MPA. To evaluate the performance of the MPASCA algorithm, extensive experiments were carried out using 18 UCI datasets. More so, the metabolomics dataset is used to test the proposed method as a real-world application. Furthermore, we implemented extensive comparisons to several state-of-art methods to verify the efficiency of the MPASCA. The evaluation outcomes showed that the MPASCA has significant performance, and it outperforms the compared methods in terms of classification measures.

Automated summary

Feature selection is an important process to reduce the dimensionality of datasets. This study proposes a feature selection method based on a modified version of the marine predators algorithm, and extensive experiments show that it outperforms other methods in terms of classification performance.