The Prediction of Buzz in Social Internet of Things

Social Internet of Things is a critical digital tool for companies of different sizes. Good governance of the Social Internet of Things allows successful strategy to emphasise brands, making products no longer discretionary. Buzz marketing is a powerful advertising strategy that focuses on exploiting the value of a promotion or item, through conversations between family members and friends, or through wider conversations on the Social Internet of Things platforms. Buzz's forecast allows for an overview of the ranking of retailers by posts shared by future consumers or members on social media platforms. Buzz estimation in the Social Internet of Things Platforms, like Twitter, is a daunting challenge over real-time data by defining multiple aspects for buzz exploration. The bulk of buzz predictor analysis studies are modelled on machine learning methods like Radial Base Feature, Neural Network Models, Ant Colony Optimization, and Support Vector Devices. In this study, Random Forest is being used to make buzz predictions as it creates multiple decision trees and mixes them together again to make predictions more precise. The findings of comparative studies with Support Vector Machine (SVM) in three separate kernels and Radial Basis Function (RBF), it suggest that Random Forest is better at forecasting buzz and has also established a rating list of important features in order to achieve reasonable prediction performance. The features suggest that the main important characteristics to define the subject being generated are either “Buzz” or “Not Buzz”. Random Forest showed an average improvement in accuracy with a rate of 99% and the highest training time. However, although the Radial Based Function method has proved to be quite efficient, the Support Vector Machine solution is of mediocre accuracy.