Pemilihan Algoritma Terbaik Untuk Klasifikasi Jenis E-Mail dengan Metode TF-IDF

Fitria, Denisa
Cahyana, Yana
Sulistya, Dwi
Baihaqi, Kiki Ahmad

Jurasik (Jurnal Riset Sistem Informasi dan Teknik Informatika)
Vol 9, No 1 (2024): Edisi Februari

Spam emails, sent en masse to numerous addresses, are a major annoyance. To combat this, effective filters are necessary, such as classification to separate spam from non-spam. This can be achieved through an anti-spam model utilizing text mining like TF-IDF. Using the KDD process, a study analyzed a dataset of 6046 entries, split 77.2% non-spam and 22.8% spam. Logistic Regression showed the best accuracy at 98%, outperforming Decision Tree (59%) and Support Vector Machine (95%). Thus, Logistic Regression emerged as the optimal algorithm for email classification.