Abstract : Hardware, software, and the recent technologies have resulted in the interconnection of sensory devices providing large amount of data. The internet conncected devices are bound to grow in time. This Internet of Things (IoT) extended in various applications like agriculture, transportation, medical, education etc is generating big data. These IoT devices are less immune to malicious network traffic and attacks. Implementation of security measures like encryption, authentication etc are seeming insufficient. Even though many efforts have been put in to produce master datasets which includes all possible attacks on IoT devices, it is still observed that many attacks are unexplored. This paper aims to validate one of the large scale, real time dataset CICIoT2023 generated using 105 devices consisting of seven types of attacks in addition to benign data. Applying machine learning algorithms such as SVM, KNN, Decision Tree, MLP and RF showcases significant improvements when the number of classes considered is 6 compared to other conventional classifications with 8 classes. Accuracy, F-Measure, precision are used as performance parameters to evaluate the dataset using machine learning algorithms.