SMS Spam Detection Using Machine Learning

Abstract

The global question of marketing mail by way of the Short Message Service is a major concern for those the one-use movable phones. In an exertion to find answers, a excess of deep education and machine learning methods have happened used. In the research, four specific algorithms—RVM, SVM, Naive Bayes, and KNN—are linked utilizing the bagging method. The results from each invention are therefore linked utilizing a adulthood-vote pattern to accomplish the final forecast. In light of the significance of correctly labelling and categorising unsolicited call SMS ideas, this item presents research on a comparative test of various passage categorization means. After the dataset is pre-treated, it is vectorised utilizing the TF-IDF approach, which prioritises exceptional conversation over average one. Achieving the greatest presentation on this data with an F1 score of 0.975176 is the Relevance Vector Machine implementation. The investigation confirmed that the proposed RVM model could effectively classify SMS spam mail and be used in real-world scenarios.