Enhanced Machine Learning Approach for Detecting Spammers in Industrial Mobile Cloud Computing Environment

This research study proposes an innovative machine learning algorithm, Spammer Identification Novel K-Means Extension (SINKEX), to fortify IoT mobile networks against commands and potential disruptions in industrial production. Utilizing an industrial mobile cloud dataset, the algorithm employs basic, content, and network features, including follower count, post frequency, message content, and hyperlink usage, for spam detection. After employing Pearson Correlation and Euclidean Distance to eliminate dissimilar records, Principal Component Analysis (PCA) is applied to remove irrelevant features. The SINKEX algorithm then labels unclassified records, designating an account as a SPAMMER or NON-SPAMMER based on features like follower count and post frequency. Comparisons with a hybrid fuzzy c-means clustering algorithm demonstrate SINKEX's superior performance, achieving higher accuracy, precision, and recall. The results are presented through an intuitive graphical user interface, showcasing the algorithm's effectiveness in securing industrial IoT networks against malicious activities.