Enhanced SMS Spam Detection Using Boost Light Approach

Mobile Telecom subscriber’s are interrupted by these spams and it becomes necessary for the service providers to take steps to protect the user. In this project, we aim toward developing an end to end SMS spam detection model using different ML algorithms available in today’s day and age. Instead, it provides latest algorithms used in group such as CatBoost, LightGBM and XGBoost to enumerate text. Using these algorithms, the system is able to very accurately classify an SMS message as spam or not. In this project each of the algorithms are tested and they are compared based on some of the key metrics (accuracy, precision, recall and F1 score). Since there is a need to apply the textual pre processing methods like TF-IDF and bag of words methods in order for the model to learn the appropriate underlying on sms message content. For this spam sms classification issue, in this project, we will attempt to make a decision for what will be the most appropriate choice for a scalable solution for the spam problem. Therefore, this will help with their security and give them a quick means to organize their communication with this approach rather than through other systems form of spam filtration.