Password Classification Using Machine Learning and Natural Language Processing Techniques: Methods and Evaluations

Passwords are the most widely used authentication method and play a crucial role in the field of information security. In this study, we explore the effectiveness of applying machine learning (ML) and natural language processing (NLP) techniques to password classification. We compare the performance of classifiers by using eight ML techniques and four NLP techniques to classify user-created passwords. The experimental results show that the classifier using a combination of Bag-of-Words and Logistic Regression outperforms other classifiers, achieving an accuracy of 98.53% and a recall for weak passwords of 99.68%.