In this paper we propose a strategy based on past/present values provided by each sensor of a network for detecting their malicious activity. Basically, we will compare at each moment the sensor's output with its estimated value computed by an autoregressive predictor. In case the difference between the two values is higher then a chosen threshold, the sensor node becomes suspicious and a decision block is activated.

Being often deployed in remote or hostile environments, wireless sensor networks are vulnerable to various types of security attacks. A possible solution to reduce the security risks is to use directional antennas instead of omnidirectional ones or in conjunction with them. Due to their increased complexity, higher costs and larger sizes, directional antennas are not traditionally used in wireless sensor networks, but recent technology trends may support this method. This paper surveys existing state of the art approaches in the field, offering a broad perspective of the future use of directional antennas in mitigating security risks, together with new challenges and open research issues.

Because of numerous possible causes involved, it isn't easy for general physicians to identify the precise reason of the psychiatric diseases and to decide the correct treatment. Bayesian networks are recognized as efficient graphical models with significant capabilities for investigating biomedical data either to obtain relationships between variables, either for medical predictions. Our paper provides a Bayesian network-based analysis of psychiatric patient data, which have been gathered from a Romanian specialized clinic during a couple of years. The development of this Bayesian network led us to the identification of most significant factors that affect some important diseases and their correlations.