Evaluating Spam Detection Techniques: A Comparison of TF-IDF and Sentence Embeddings with Machine Learning Models

This paper presents a comparative analysis of spam detection models using two feature extraction techniques: TF-IDF and Sentence Embeddings. The study evaluates three machine learning models: Random Forest, XGBoost, and Logistic Regression, to classify spam messages. Optuna, a hyperparameter optimization framework, is applied to tune Random Forest and XGBoost classifiers. Results are compared in terms of accuracy, precision, recall, fl-score, support, execution time, train time and tune time.