SASSO: Design of a Signature-Based Feature Selection Model with Ensemble Deep Learning Using Bi-LSTM and Bi-GRU for Chronic Kidney Disease Classification
Year: 2024
Type: article
Abstract: Chronic kidney disease (CKD) involves numerous variables, but only a few are significant for classification. The SES method, inspired by constraint-based learning in Bayesian networks, identifies essential features in CKD. Unlike traditional feature selection methods, which concentrate on a single set of features with the greatest predictive potential, the SES method can identify multiple predictive feature subsets with comparable performance. Most feature selection (FS) classifiers work better with strongly correlated data, making FS challenging in high-throughput data for finding important features and choosing the best classifier. This study is conducted on a real time data available from hospitals. This study suggests the use of the Least Absolute Shrinkage and Selection Operator (LASSO) in conjunction with the SES method, which is abbreviated as SASSO, to identify CKD features. Subsequently, a combination of Bidirectional Long Short-Term Memory (Bi-LSTM) and Bidirectional Gated Recurrent Unit (Bi-GRU) ensemble deep-learning models is proposed for CKD classification. The model's performance is measured using accuracy, precision, recall, and the F1-score. The experimental results are compared to individual classifiers, including Random Forest (RF), Naïve Bayes (NB), XGBoost (XGB), and Artificial Neural Networks (ANN). The findings show a 6% improvement in classification accuracy with the proposed hybrid feature selection approach and the Bi-LSTM-Bi-GRU ensemble model.
Author: Anindita Khade
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7217463552
