Chronic Kidney Disease Prediction Based On Machine Learning Algorithms

^1,2,3 Department of CSE, Vignan’s Foundation for Science, Technology and Research Vadlamudi, Guntur, Andhra Pradesh, India
⁴ Department of Computer Science and Engineering, Vignan’s Foundation for Science, Technology and Research, Vadlamudi, Guntur, Andhra Pradesh, India

¹ Email: kethinenilikitha005@gmail.com

Abstract

Chronic kidney disease is a progressively bad disease. In most cases, there are serious complications with the kidneys if early identified. Early prediction and diagnosis of CKD are quite important to improve the management and effectiveness of the treatment. This paper analyses the suitability of several machine learning algorithms for CKD prediction in consideration of health-related parameters like age, blood pressure, and blood glucose levels. An integrated dataset of 400 patients has been used for this study, which comprises 24 health- It was performed by employing a variety of classifiers including DT, KNN, RF, SVM ,CNN and Hybrid. For the performance of each model, the accuracy, precision, recall, and F1-score metrics were employed to check its diagnostic ability. It was found that the maximum accuracy was achieved by the Random Forest model that justified its applicability as an effective tool for the early diagnosis of CKD. This study is the perfect example of the use of machine learning techniques in the field of medical diagnoses, with their contribution to the development of future health care. Several recommendation methods were used: KNN Basic, Nonnegative Matrix Factorization (NMF), Co-Clustering, and Singular Value Decomposition (SVD). KNN Basic model obtained a root mean square error of 0.5007 with an accuracy of 44.50. The NMF model gave a better result than the first one with RMSE 0.4999 and accuracy 51.50. Co-Clustering resulted in an RMSE of 0.5000 and an accuracy of 50.50. Most notably, the last SVD model greatly Identify applicable funding agency here. If none, delete this. performed compared to the rest, achieving an RMSE of 0.2261 and an outstanding accuracy of 90.40, highlighting the an incredibly powerful tool for a more personalized type of care. Providers would then have the ability to identify ahead of time high-risk people and monitor their care more carefully. It, however, opens up the possibility of unsupervised learning in health analytics and shows how this can be applied to the role of machine learning for early detection and targeted treatment, thereby contributing to better patient outcomes and proactivity in managing heart disease risks. necessity of more sophisticated methods in recommendation systems. These results not only show how different recommendation algorithms work but also point out the significant influence of hyper parameter tuning on improved predictive precision. This work provides essential contributions to the knowledge on the effectiveness of customized recommendations, which suggests insight for future developments in e-commerce strategies.