Enhancing the security of patients' portals and websites by detecting malicious web crawlers using machine learning techniques.

•Feature selection based on navigational behavior of users of web sites and patient portal is very important to detect malicious crawlers effectively.•The new features which is proposed in this study, “Unique pages” and “Hit”, are very effective in separating usual users from crawlers. Especially, “Unique pages” has high correlation with class feature and rank second in importance factor.•Support vector machine technique has a very good performance among the machine learning techniques to separate the crawler visitor and human users of websites.