Rumor Detection in Business Reviews Using Supervised Machine Learning

Currently, a high volume of business data is generating with a high velocity in different forms like unstructured, structured or semi-structured. Due to social media arrival, there is a deluge of business rumors and their manual screening is time-consuming and difficult. In the current social computing era, it is necessary to move towards an automated process for the detection of business rumors. This work aims at developing an automated system for detecting business rumors from online business reviews using supervised machine learning classifiers, namely Logistic Regression, Support Vector Classifier (SVC), Naïve Bayesian (NB), K-Nearest Neighbors (KNN) to classify the business reviews into rumor and nonrumor. Experimental results show that Naïve Bayesian (NB), achieved efficient results with respect to other classifiers with an accuracy of 72.43 %.