Wireless Sensor Networks (WSN) has gained attention as it has been useful in warning about disasters. Predicting natural disasters like hailstorm, fire, rainfall etc. by WSN are infrequent and stochastic. This is an important topic of research. Detection of these disasters should be fast and accurate as they may cause damage and destruction at a large scale. In this paper, comparison of various machine learning techniques such as SVM, regression, decision trees, neural networks etc. has been done for prediction of forest fires. The proposed approach in this paper presents how regression works best for detection of forest fires with high accuracy by dividing the dataset. Fast detection of forest fires is done in this paper by taking less time as compared to other machine learning techniques.

Efficient Approach for Securing Message Communication in Wireless Sensor Networks from Node Clone Attack

Wireless sensor networks (WSN) consist of small size sensor nodes with limited sensing, processing, communication, and storage capabilities. These sensor nodes are vulnerable to the node clone attack where the attacker compromises the node and extracts secret information from the node and replicate the large numbers of clones of captured node throughout the sensor network. Therefore, providing security in such networks is of utmost importance. The main challenge to achieve this is to make the security solution energy efficient so that it is feasible to implement in such resource constrained nodes in WSN. In this paper, an energy efficient algorithm is proposed for node authentication. The aim of node authentication algorithm is to authenticate the sensor nodes before message communication within WSN so that cloned nodes are identified in the initial step of the communication. This algorithm uses encryption decryption operations and also XOR, extraction, bitwise shift operations. The performance of the proposed algorithm is analysed in terms of communication, storage, and computation overheads metrics. Finally, performance of the proposed algorithm is analysed with the other node authentication algorithms.