Network monitoring and analysis have taken a primordial role in understanding the functioning of the internet and preventing it from any future probable attacks. Many publications treated the topic of security as an interesting research domain. Due to the importance and the different applications of wireless sensor networks, the research of finding security solutions is always up but their limited resources (energy consumption, CPU capacity, limited memory…) make the application of classical protocols of security like cryptography, almost impossible. Therefore, recently machine learning is a new field that has gotten the interest of the majority of researchers. The solutions for providing security services conveyable in this type of network are proposed. In this paper, we focus on the monitoring of wireless sensor networks using machine learning as an intrusion detection system. We address in particular active attacks. Because of the limitation of machine learning and to provide a high protection against attack we proposed a hash cryptographic function to protect the authenticity of information. Our interesting contribution aims at three important axes of security: confidentiality, integrity, and authentication.