Prediction of Marine Water Quality Index Using a Stacked Classifier Under Machine Learning Architecture

Abstract

The health of humankind is intrinsically associated with the health of the marine and ocean ecosystems. The pollution of the coastal region due to urbanization, for example, principally harms the growth of the ecosystem with poor-quality of water, which aggravates the survival of marine organisms and animals. The toxicity of the contaminated seafood would affect the human-ocean ecosystem thereby bringing down the economic rank of the region as well. Therefore, it is mandatory to assess the quality of the marine and ocean water to initiate any statutory measures to protect the regional marine water against pollution and dumping of toxic matter. This paper, therefore, presented an architecture of machine learning techniques to assist in classifying marine water quality. The proposed framework evaluated various classification models and selected the best fit out of the top-performing algorithms through training and optimizing. The finalized model was a stacked classifier, which was then deployed to predict the marine water quality index from the physicochemical and biological properties of the water.

Keywords

• machine learning, model training, model optimizing, ensemble learning, marine water quality, marine water quality index