Secure water quality prediction system using machine learning and blockchain technologies

Water is important for every organism, especially human survival. 2-3 % of fresh water is available on the earth's surface. Discharge of contaminated municipal sewage, removal of degradable wastes and industrial effluents has polluted freshwater resources like an ocean, river, pond, channel, or lake. Hence, this precious resource must be carefully maintained and preserved before consumption. In this research, machine learning models such as Linear Regression, Generalized Linear Model, Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), classification and regression trees, and Random Forest were used to predict the water quality parameter of Chittar Pattanam Channel, Kanyakumari district, Tamil Nadu in India by giving latitude and longitude. The results showed that the Random Forest (RF) algorithm was better than other models in terms of prediction accuracy with a mean absolute error of 0.56, mean square error of 0.33, and root mean square error of 0.56. Blockchain technologies were used to provide security in the machine learning model. In this work, more than one authorized person is involved in the prediction process, and the authorized person is verified by his signature using Secure Hash Algorithm-256 (SHA). To generate an unpredictable and unique key, SHA-2 uses the size of hash values is 256,384 and 512, a message size is 1024, total rounds are 80 and a word size is 64bits. RSA (Rivest-Shamir-Adleman) technique is used for performing data transfer of keys and encrypting and decrypting data. This study implements a secure water quality prediction system to reduce pollution and improve water quality.

Automated summary

This article discusses the importance of water and the pollution of freshwater resources, and introduces the use of machine learning models and blockchain technology to predict and protect water quality.