A Deep Neural Network for Early Detection and Prediction of Chronic Kidney Disease

Area of Science:

• Nephrology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Chronic Kidney Disease (CKD) is primarily caused by diabetes and high blood pressure.
• Early detection of CKD is crucial for preventing disease progression and improving patient outcomes.
• Current diagnostic methods face challenges in identifying various CKD-related diseases early.

Purpose of the Study:

• To develop a novel deep learning model for the early detection and prediction of CKD.
• To compare the performance of the proposed deep neural network against contemporary machine learning techniques.
• To identify key features indicative of CKD for improved diagnostic accuracy.

Main Methods:

• A deep neural network was developed and its parameters optimized through multiple trials.
• Missing values in the dataset were imputed using the average of associated features.
• Recursive Feature Elimination (RFE) was employed to select the most significant features.
• Key features identified include Hemoglobin, Specific Gravity, Serum Creatinine, Red Blood Cell Count, Albumin, Packed Cell Volume, and Hypertension.

Main Results:

• The proposed deep neural network model achieved 100% accuracy in CKD classification.
• The deep learning model significantly outperformed Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Logistic Regression, Random Forest, and Naive Bayes classifiers.
• Recursive Feature Elimination identified critical biomarkers for CKD detection.

Conclusions:

• The developed deep learning model demonstrates superior performance for early CKD detection.
• This novel approach offers a potentially valuable tool for nephrologists in diagnosing CKD.
• Accurate early detection using AI can lead to timely interventions and better patient management.