Wireless sensor networks are vulnerable to many types of security attacks, including false data injection, data forgery, and eaves dropping. Sensor nodes can be compromised by intruders, and the compromised nodes can distort data integrity by injecting false data. The transmission of false data depletes the constrained battery power and degrades the band-width utilization. This paper examines various techniques to detect false nodes i.e. outliers. The main emphasis is to detect outliers on the basis of distance measures. Thus clustering and support vector machines are used as a basis.