FTDT: Rough set integrated functional tangent decision tree for finding the status of aqua pond in aquaculture

Abstract

Due to the increasing demand on aquaculture, the continuous monitoring of water quality and characteristics are significant for maximizing the yields. Even though many physicochemical parameters are available for monitoring the water quality, the knowledge of domain experts are expected to analyze these parameters to find the final decision about the quality of water. In order to utilize the knowledge of the domain experts for water quality artificially, we have developed a functional tangent decision tree algorithm which, predict the water quality of the ponds based on the physiochemical parameters. The proposed method of predicting the water quality consists of three important steps such as, uncertainty handling, feature selection using reduct and core analysis, classification using the functional tangent decision tree. The proposed functional tangent decision tree is constructed by utilizing a new function called functional tangent entropy for the selection of attributes and split points. Then, the proposed work is experimented with the real database and compared with the existing classifiers to prove its better performance. The proposed work obtained the maximum classification accuracy of 95% as compared with the existing work.