AI-Assisted Pair Programming System for Automated Code Analysis, Refactoring, and Explanation

Abstract

Water quality assessment is a critical aspect of environmental monitoring and public health management. With increasing industrialization, urbanization, and agricultural activities, water bodies are continuously exposed to pollutants, making it essential to monitor and evaluate water quality efficiently. Traditional methods of water quality analysis rely on laboratory testing, which is time-consuming, expensive, and not suitable for real-time monitoring. To overcome these limitations, this research proposes an intelligent water quality prediction system using machine learning techniques for potability assessment.The proposed system leverages data-driven approaches to classify water samples as potable or non-potable based on physicochemical parameters. The implementation is developed using Python and integrates machine learning algorithms such as Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Networks (ANN). A graphical user interface is designed using a GUI framework to facilitate user interaction, enabling easy dataset upload, preprocessing, model execution, and result visualization.The system begins with data acquisition, where water quality datasets containing parameters such as pH, hardness, solids, chloramines, sulfate, conductivity, organic carbon, trihalomethanes, and turbidity are used. Preprocessing techniques are applied to handle missing values and normalize the dataset. The data is then divided into training and testing sets to evaluate model performance.Feature extraction and encoding techniques are used to transform the dataset into a suitable format for machine learning models. The SVM and Random Forest algorithms are employed for classification, providing robust performance for structured data. Additionally, an Artificial Neural Network is implemented with multiple hidden layers to capture complex nonlinear relationships among water quality parameters. The ANN model is trained over multiple epochs, and its performance is evaluated using accuracy metrics.The system also includes a visualization module that plots accuracy and loss graphs, providing insights into model performance during training. The prediction module allows users to input new data and obtain real-time classification results, indicating whether the water is safe for consumption.Experimental results demonstrate that the proposed system achieves high classification accuracy and effectively distinguishes between potable and non-potable water samples. Among the implemented models, the neural network shows improved performance in capturing complex relationships, while Random Forest provides stable and reliable results. The proposed framework offers a scalable, efficient, and cost-effective solution for water quality prediction. It reduces dependency on manual testing and enables rapid decisionmaking. This system can be extended for real-time monitoring using IoT devices and integrated into smart environmental management systems. The research contributes to the development of intelligent water quality assessment tools, promoting sustainable resource management and public health protection.