The exponential expansion of the Internet of Things (IoT) has resulted in a heightened level of automation and connectedness among gadgets, eliminating the need for deliberate user involvement, thereby augmenting the overall quality of our lives. Nevertheless, the security of IoT devices is a paramount issue since they are susceptible to cyber-attacks, which can result in substantial harm if not promptly identified and addressed. The present work introduces a methodology that combines three machine learning classifiers with principal component analysis (PCA) to efficiently and precisely identify IoT network attacks. The suggested strategy is evaluated using the Bot-IoT dataset, and the observed findings demonstrate substantial efficiency in detection performance when compared to current methodologies that use the whole features vector. In the evaluation, the DNN algorithm achieved an 88.6% f1-score, while the SVM and Decision tree reached 87.21% and 80.58% respectively. In contrast, the accuracy of the other ML algorithms did not fall below 95%. The proposed methodology can be further expanded to augment the security of other Internet of Things (IoT) applications, therefore presenting a prospective addition to the domain of IoT security.