Safeguarding SMS: A Dynamic Duo Approach to Tackle Spam Using LDA and QDA

In our daily lives, the usage of mobile phones is increasing day by day. This trend has become more advantageous for scammers, who use mobiles as an intermediary device for their scams on individuals. Among various types of scams, one common method is sending spam SMS messages to people. These messages often contain fake links and aim to acquire personal details such as credit card numbers, as well as promoting various products. To combat these spam SMS messages, a spam detector has been implemented. The detector employs classification models, specifically Linear Discriminant Analysis and Quadratic Discriminant Analysis. The Spam Collection dataset serves as the dataset for this project. The dataset has been modified for further processing. Additionally, various Natural Language Processing (NLP) techniques, including tokenization, handling of stop words, applying regular expressions, and Porter stemming, have been applied to the dataset. The outcomes of these NLP techniques have been converted into a corpus, which is used for count vectorization.For the classification, both Linear Discriminant Analysis and Quadratic Discriminant Analysis models have been used. The accuracy, precision, recall, F1-score, confusion matrix, and ROC curve have been estimated for both models. A comparison of the performance of these models has been made based on accuracy, precision, recall, F1-score, confusion matrix, and the ROC curve.