Globally, billions of people continue to suffer from the contamination of water bodies with its scale and degree of impact ever-increasing. Yet, compared to the hundreds of thousands of rivers existing worldwide, the extent of studies done to analyze the water quality of rivers seem miniscule. As industrialization continues to predominantly contaminate the surface water of river flows, examination of water bodies becomes increasingly dire. This research compares and evaluates the industrial pollution within the Han River of South Korea and the Donau River of Austria, interpreting the independently collected data of water parameters of dissolved oxygen (DO), pH and total dissolved solids (TDS) as well as reviewing existing information regarding the two rivers, both quantitative and qualitative. Public data human population density (HPD) and air quality index (AQI) was utilized. Using cumulative air quality index (CAI) as the AQI for both rivers, All data revealed the exacerbating level of industrial pollution within the Han River in contrast to the Donau River experiencing manageable degree of contamination. The data were used to create matrices of Pearson’s correlation coefficients as well, of which the matrix cumulatively evaluating the data collected showed strong correlation between all pairs of the parameters examined, further validating the evaluation of the industrial pollution among both rivers. The data was used for training and testing 185 combinations of supervised machine-learning models in Python, applying various layers of feature selection. Ultimately, data from public databases were gathered to create linear models, nonlinear support vector regressor(SVR) models and model ensembles, applying k-fold Cross Validation while distinguishing between feature selection methods and dimensionality reduction techniques to ensure high accuracy.
 
Keywords: environmental science, ecology, river hydrology, industrial pollution, machine learning