This paper presents a feature selection method based on rough sets and particle swarm optimization (PSO). The proposed method, called PSORSFS, combines the strengths of rough sets for feature selection with the global search capability of PSO. Rough sets are used to identify minimal reducts, which are subsets of features that preserve the classification ability of the original feature set. PSO is employed to efficiently search for optimal feature subsets by simulating the movement of particles in a problem space. The PSORSFS algorithm represents each particle as a binary string, where each bit indicates whether a feature is selected. The velocity of each particle determines how many features are changed in each iteration to approach the best solution. The algorithm uses a fitness function that balances classification accuracy and feature subset length. The fitness function is defined as α * γ_R(D) + β * (|C| - |R|)/|C|, where γ_R(D) is the classification quality, |R| is the number of selected features, |C| is the total number of features, and α and β are parameters that control the importance of classification accuracy and subset length. The algorithm is tested on 27 UCI datasets, and compared with other feature selection methods such as POSAR, CEAR, DISMAR, and GAAR. The results show that PSORSFS outperforms these methods in terms of classification accuracy and computational efficiency. PSORSFS is able to find optimal or near-optimal feature subsets quickly, even for large datasets. The algorithm is also more efficient in terms of computational time compared to other methods, especially for datasets with a large number of features. The PSORSFS algorithm is effective in finding minimal reducts, which are subsets of features that preserve the classification ability of the original feature set. The algorithm is able to find optimal solutions in a relatively small number of iterations, making it suitable for real-world applications where computational resources are limited. The use of PSO allows the algorithm to explore the feature space efficiently and find good solutions without getting stuck in local optima. The algorithm is also able to handle noisy and irrelevant features, making it suitable for real-world data analysis. The results show that PSORSFS is a promising method for feature selection in rough set-based applications.This paper presents a feature selection method based on rough sets and particle swarm optimization (PSO). The proposed method, called PSORSFS, combines the strengths of rough sets for feature selection with the global search capability of PSO. Rough sets are used to identify minimal reducts, which are subsets of features that preserve the classification ability of the original feature set. PSO is employed to efficiently search for optimal feature subsets by simulating the movement of particles in a problem space. The PSORSFS algorithm represents each particle as a binary string, where each bit indicates whether a feature is selected. The velocity of each particle determines how many features are changed in each iteration to approach the best solution. The algorithm uses a fitness function that balances classification accuracy and feature subset length. The fitness function is defined as α * γ_R(D) + β * (|C| - |R|)/|C|, where γ_R(D) is the classification quality, |R| is the number of selected features, |C| is the total number of features, and α and β are parameters that control the importance of classification accuracy and subset length. The algorithm is tested on 27 UCI datasets, and compared with other feature selection methods such as POSAR, CEAR, DISMAR, and GAAR. The results show that PSORSFS outperforms these methods in terms of classification accuracy and computational efficiency. PSORSFS is able to find optimal or near-optimal feature subsets quickly, even for large datasets. The algorithm is also more efficient in terms of computational time compared to other methods, especially for datasets with a large number of features. The PSORSFS algorithm is effective in finding minimal reducts, which are subsets of features that preserve the classification ability of the original feature set. The algorithm is able to find optimal solutions in a relatively small number of iterations, making it suitable for real-world applications where computational resources are limited. The use of PSO allows the algorithm to explore the feature space efficiently and find good solutions without getting stuck in local optima. The algorithm is also able to handle noisy and irrelevant features, making it suitable for real-world data analysis. The results show that PSORSFS is a promising method for feature selection in rough set-based applications.