Classification of borehole waters in terms of potability by machine learning techniques

Water is essential for the survival and sustenance of all living things. Therefore, it is very important to maintain the balance of water quality. Otherwise, it can cause serious harm to human health and seriously affect the ecological balance among other species. Water quality is an important factor to consider both in terms of ecosystem needs and pollution levels that directly affect health, hygiene, food and economy. Increasing global water demand is leading to over-exploitation of groundwater resources and degradation of groundwater quality. Physical and chemical properties, which are significantly influenced by geological formations and human activities, indicate how groundwater quality is changing. Accurate and reliable assessment of groundwater resource information is an important element for effective management of groundwater quality. Thesis study, Machine Learning (ML) classification algorithms, an effective artificial intelligence technique for predicting water quality as an alternative to traditional analysis methods in water monitoring applications, were used to classify artesian waters in terms of potability. In this context, 30 potable and 30 non-potable labeled borehole water data obtained from Special Provincial Administration of Elâzığ were analyzed with Decision Trees (DT), K Nearest Neighbor (KNN), Naive Bayes (NB), Random Forest (RF) and Support Vector Machines (SVM) algorithms. 80% of the dataset was used for training ML models and the remaining 20% was used for testing. As a result of the study, the KNN algorithm showed the best classification performance with an accuracy of 99.93%.