Due to the unmanned nature of Wireless Sensor Networks, security becomes a key criterion when it comes to networks dealing with confidential data. Compromised node, Denial of Service (DoS) [1] attacks and Black-Holes/Sink-Holes [2] are the three key types of attacks in Sensor Networks. Classic routing algorithms use deterministic multipath routing schemes, where a predefined path exits between any two nodes. Once if the adversary acquires the routing algorithm it is possible to compute the route, making all information sent over these routes vulnerable to its attacks. Our approach involves selecting intermediary nodes for each packet rather than sending the packets directly to the destination node. This way, the user initially disperses all the packets that are to be transmitted using a modified form of Backpressure algorithm [4] and then directs them to the destination node using SENCAST [5]. By following this method, most of the packets that are sent through a network have the probability of escaping black holes. Simulations show that our approach is much more effective in terms of security when compared to their deterministic counterparts.

The development of many educational institutions is based on the performance of students learning and understanding capabilities. Here, we analyzed their academic profile with their grades and various cumulative attributes. The academic performance in learning their subjects could be improved by motivational approach. The analysis of student performance is carried out through knowledge-based data mining process. But, the problem is arrived by a probability of information prediction accuracy from student data set which is not accurate. Here, we propose a novel machine learning algorithm based on subspace clustering and multi-perspective classification techniques to identify psychological motivation required students. Also, the extraction of relational patterns to form enhanced clustering classes is done. This discovers the innovative relations between students and their educational performance in the various attributes using surf scale nested clustering approach based on an intelligent predicting system from soft computing processing tasks. This improves the data prediction rate by considering the time factor analysis and complexity to design and develop an efficient clustering algorithm which maximizes the clustering and classification accuracy for improving academic performance.

Network Applications can be broadly classified as Throughput sensitive or Delay sensitive. Such applications require efficient routing mechanisms in order to work effectively. Genetic Algorithms can be used for defining the Best or Optimal Route based on its sensitivity and Various Constraints that the application imposes usually referred to as Constraint Satisfaction Problems (CSP).The use of Genetic Algorithms for selecting an optimal route based on CSP requires a mechanism for automatic discovery of network topology and also mechanisms for learning the capacity of the network infrastructure. SNMP along with MIB provides the required data for Topology discovery and also to learn the information about the current network capacity along with various bottlenecks present in the topology. Since the shortest path is not always the best path, our genetic algorithm will provide the optimal route based on CSP and application sensitivity.

No abstract

Now-a-days