Chronic Kidney Disease (CKD) Prediction using AutoDL Classifier Algorithms

Background: Chronic kidney disease (CKD) is a health problem that is getting more common and harder to treat. It has an effect on many people. It is important to be able to predict how chronic kidney disease will progress so that doctors can give personalized care and intervene early. The models that are currently available don't always accurately predict how well CKD will progress because they don't always take into account the factors that change over time. Methodology: This study shows a new way to handle categorical variables and overfitting while also getting the time-based information that is needed to make accurate CKD predictions. Findings and Results: We used Automated Deep Learning (AutoDL) with Convolutional Long Short-Term Memory (convLSTM) for feature extraction, Attention Long Short-Term Memory (LSTM) for classification, and Bidirectional Long Short-Term Memory (biLSTM) for temporal prediction to deal with the problems that come with predicting CKD during this study. The proposed method is better than the best models in a number of important areas, such as accuracy, recall, and precision.