Exploring the Impact of IoT Data Diversity on Machine Learning Model
Dhananjay Bisen, Nookala Venu, Aditya Dubey, Priyanka Garg, Suresh Salendra, ThIbetombi Devi · 2024
The past few years have witnessed a remarkable change by the rapid expansion of the Internet of Things (IoT). The increasing spread of networked devices in various areas of the Internet of Things has resulted in a huge amount of machine data. This data contains rich information that can be harnessed across domains such as healthcare, agriculture, smart cities, and manufacturing. To effectively analyze and capitalize this information, the integration of machine learning (ML) methods is essential. Thus, well-known machine learning techniques including Logistic Regression, Decision Tree, KNN, SVM, Random Forest, and Naive Bayes are used in paper. The performance of the algorithms is analyzed using three diverse IoT datasets and compared to one another using pre-defined measures such as accuracy, precision, recall, F1-score and execution time. Furthermore, test findings shown that the random forest algorithm consistently outperformed other algorithms for a variety of datasets with up to 98% accuracy.