Paper Title
Evaluating The Efficacy of Machine Learning Techniques in Chronic Renal Insufficiency

Abstract
Chronic Renal Insufficiency (CRI) is a serious health condition that involves the progressive deterioration of kidney function over an extended period, often with minimal visible symptoms. This condition is increasingly recognized as a global health challenge, exacerbated by factors such as unhealthy eating habits and insufficient hydration. Timely detection is crucial for effective management, and machine learning (ML) techniques have demonstrated substantial effectiveness in predicting CRI at early stages. This study presents a comprehensive a method for predicting Chronic Renal Insufficiency (CRI) with clinical data. Data preprocessing, feature selection, model construction, performance evaluation, and deployment via a Flask-based application are some of the phases in the process. To create unique prediction models, a range of Machine learning techniques like K-Nearest-Neighbors and Naïve-Bayes were utilised. Linear-Discriminant-Analysis, Light-GBM, AdaBoost, XG-Boost, Logistic-Regression, and Multi-Layer Perceptron. By employing ensemble techniques, the study achieved a remarkable accuracy rate of approximately 99%. Utilizing a publicly available CRI dataset, the models demonstrated superior performance relative to previous research, indicating that this approach provides more dependable predictions. Additionally, the system offers treatment suggestions for patients with positive predictions, facilitating early intervention and effective management. The incorporation of a Flask-based interface ensures user-friendly deployment, making the predictive tool readily accessible to healthcare providers and patients, thereby enhancing its practical applicability in clinical environments. Finally, the research underscores the necessity for ongoing monitoring and follow-up for patients identified as at risk, ensuring that prompt interventions and lifestyle changes can be implemented to mitigate disease progression and enhance patient outcomes. Keywords - Multi-Layer Perceptron (MLP), Ensemble techniques, Accuracy, Light-GBM, X-GBoost, Disease management, machine learning, clinical data.