DFTDT: distributed functional tangent decision tree for aqua status prediction in wireless sensor networks

With the inspiration of applicability of sensor nodes in various applications, such as wildlife monitoring, military target tracking and surveillance, hazardous environment exploration, and natural disaster relief, the continuous monitoring of water quality and characteristics can also be a significant application to monitor the physicochemical parameters for maximizing the yields. Accordingly, a variety of sensors can be located inside the ponds to collect the required parameters and the detection of water quality can be done using the data classification algorithms. In this paper, we have proposed a distributed functional tangent decision tree (DFTDT) classifier to predict the quality of water in wireless sensor networks. At first, the wireless sensor node is used to sense the data from the pond and the functional tangent decision tree is constructed by utilizing the functional tangent entropy for selection of attributes and split points. Here, routing path is optimally identified using cluster head-based routing protocol based on fractional calculus artificial bee colony algorithm, in which the individual decision trees are merged along the routing path. Then, the results of cluster head-based routing protocol are sent to sink node, in which the proposed DFTDT classifier is used to classify the water quality parameter using the randomly generated pseudo data. Finally, the networking performance of the proposed algorithm can be evaluated using normalized energy consumption with the existing works. From the results, we proved that, the proposed algorithm achieves the better prediction accuracy as 80%.