Efficient Multiclass Classification Using Feature Selection in High-Dimensional Datasets

Feature selection has become essential in classification problems with numerous features. This process involves removing redundant, noisy, and negatively impacting features from the dataset to enhance classifier’s performance. Some are less useful than others or do not correlate system’s evaluation, their removal does affect In most cases, a monotonically decreasing impact on performance increases accuracy. Therefore, this research aims propose dimensionality reduction method using feature technique paper proposes novel feature-selection approach that combines filter wrapper techniques select optimal Mutual Information Sequential Forward Method 10-fold cross-validation. Results show proposed algorithm can reduce by more 75% datasets large achieve maximum accuracy of 97%. The outperforms performs similarly existing ones. could be better option for minimized