Improving Requirements Classification with SMOTE-Tomek Preprocessing
The episode discusses 'Improving Requirements Classification with SMOTE-Tomek Preprocessing,' a paper by Barak Or. Hosts explain that combining SMOTE and Tomek preprocessing improves AI's ability to sort requirements data. The method simultaneously balances classes and cleans noisy boundaries, leading to sharper, more reliable models.