Performance Comparison of Machine Learning Algorithms in Short Message Service Spam Classification

With the introduction of short message service (SMS) in the second-generation mobiles, we find it being exploited by many companies. They tend to spread unwanted advertisements and offers to the end user. These messages are a huge disturbance to the customers. Often customers find it difficult to receive the desired messages. As technology is advancing many methods have been tested for preventing these spam messages from reaching the user. This has been done with the help of machine learning techniques. Some of the popular machine learning techniques to filter out spam messages from ham are logistic regression, SVM, KNN, Naïve Bayes, Decision Tree and Random Forest. This paper focuses on comparing how well all these algorithms classify the spam messages and also determine their accuracy. Based on the findings Support Vector Machine (SVM) filters the messages the best.