Distributed Nearest Neighbor-based outlier identification technique for WSNs

In this research paper, we provide a distributed K-mean strategy based on dynamic Neighbor co-ordination. It deals with locating problematic nodes by using intermittent faults and clustering mechanisms in the sensor network. The described model, sensor node, takes into account the average of all the cluster regions by using the median-based K-mean approach to gather the data from nearby sensors within the specified environment. Every region's cluster head gathers the data that has been captured from all nearby sensor nodes. Every specified interval of time, a fault occurs, and the system calculates the data's median and takes its average. It determines the separation between each cluster head and the sensor value based on the distance and average data. The proposed method is rigorously tested, with the cluster head presumed to be the trustworthy node that reliably supplies the right data. The correctness is determined after taking into account the data from the dispersed cluster heads. With regard to the different parameters are to use for predicting the accuracy of data, fault positive rate and fault alarm ratio over the data transmission. This proposed model is contrasted with other ones already in use. The outcomes of the statistical analysis show that the proposed methodology produces an accurate result as compared to the traditional or existing approaches.