Supervised machine learning, particularly classification techniques, has been used to automate and enhance the performance of intrusion detection systems (IDS). Existing frameworks for IDS focus on improving classification results through normalization, transformation, and feature selection techniques, which have positively affected the classification results. Besides these techniques, the relative significance of features has a core impact on the classification process, as it improves the quality of the utilized data. Accordingly, feature weighting is implemented using the Gini score in this paper. The weights are multiplied by the input data using matrix multiplication. The results are then used to update the training data, which is subsequently tested on multiple classifiers. As overfitting was encountered, hierarchical classification was implemented. The experiments were conducted using the KDD Cup 99 dataset, as it is challenged with noise and redundancy and widely used for evaluating various approaches in the IDS field. A k-fold cross-validation technique is used for the evaluation process, and accuracy, precision, recall, and F-measure are used as evaluation measures. The results showed that feature weighting improves the results in the multi-class classification problem with an accuracy of 99.28% compared to 98.54% using an unweighting approach. For the binary classification problem, the proposed approach achieved an accuracy of 99.7% compared to 99.65% using an unweighting approach. For precision, recall, and F-measure, the proposed approach achieved 0.9998, 0.994, and 0.9981, respectively, compared to 0.9963, 0.9998, and 0.9981 for the unweighted approach.