Forecasting of Groundwater Quality Utilizing an Efficient K-Means with Principal Component Analysis Approach

Groundwater stands as a pivotal freshwater resource in numerous regions worldwide, but the burgeoning population and diverse economic activities have adversely impacted its quality. However, the K-means algorithm is sensitive to the initial placement of centroids, where outliers can significantly influence the position of centroids resulting in suboptimal clusters. To address this, it is recommended to remove or clip outliers before applying K-means scales well to large datasets making it suitable for handling substantial amounts of data. In this proposed method, attributes are first calculated using Principal Component Analysis (PCA) determines the weight of each attribute based on eigenvalues. The PCA, by determining attribute weights and Rapid Centroid Estimation (RCE), by establishing the initial cluster centre, contribute to the overall effectiveness of the method. The classification step utilizes a Linear Support Vector Machine (SVM), suitable for linear hyper line data when the dataset can be classified into two classes using a single straight line. The results obtained showcase the superiority of the proposed method, achieving a Root Mean Squared Error (RMSE) of 1.23, R-squared (R2) of 0.90, Mean Absolute Deviation (MAD) of 0.28 and E of 0.97 on the Indian groundwater dataset. These outcomes affirm better detection and classification performance compared to other existing methods, such as Random Forest (RF), Decision Tree (DT) and K-Nearest Neighbour (KNN).