Classifying spam emails using agglomerative hierarchical clustering and a topic-based approach

Spam emails are unsolicited, annoying and sometimes harmful messages which may contain malware, phishing or hoaxes. Unlike most studies that address the design of efficient anti-spam filters, we approach spam email problem from a different novel perspective. Focusing on needs cybersecurity units, follow topic-based for addressing classification into multiple categories. We propose SPEMC-15K-E SPEMC-15K-S, two datasets with approximately 15K each in English Spanish, respectively, label them using agglomerative hierarchical clustering 11 classes. evaluate 16 pipelines, combining four text representation techniques -Term Frequency-Inverse Document Frequency (TF-IDF), Bag Words, Word2Vec BERT- classifiers: Support Vector Machine, Näive Bayes, Random Forest Logistic Regression. Experimental results show highest performance is achieved TF-IDF LR dataset, F1 score 0.953 an accuracy 94.6%, while Spanish NB yields 0.945 98.5% accuracy. Regarding processing time, leads to fastest classification, 2ms 2.2ms average, respectively.