A Hybrid Linear Discriminant Analysis and Artificial Neural Network Framework for High-Accuracy Prediction of Chronic Kidney Disease

Chronic kidney disease (CKD) is becoming an increasingly important global health issue and it is currently one of the major causes of death, predicted to become the fifth leading cause of years of life lost by 2040. It is challenging to detect it during its early stages as the disease is typically asymptomatic at this point, and nonrenal factors can interfere with the traditional diagnostic indicators. This paper describes a hybrid machine-learning approach that uses Linear Discriminant Analysis (LDA) for supervised dimensionality reduction and feature extraction and Bayesian-optimized Artificial Neural Network (ANN) for nonlinear classification. Using the publicly available UCI Chronic Kidney Disease dataset (consisting of 400 cases and 25 clinical and laboratory attributes), the method handles missing values, encodes the categorical variables, ranks the clinically relevant features, projects the data onto a maximally discriminative subspace, and classifies the patients using a regularized multilayer perceptron. The hybrid LDA–ANN model obtained an accuracy of 98.75% on the independent test set, with precision, recall, and F1-scores varying between 0.97 and 1.00, significantly beating logistic regression, decision trees, support vector machines, random forests, and standalone LDA (96.25%). The feature importance analysis showed that the selected predictors were clinically relevant features, comprising specific gravity, haemoglobin, albumin, diabetes status and hypertension. The findings show that the combination of LDA's ability to maximize class separability and the ANN's ability to model complex nonlinear interactions results in a robust, computationally efficient, and clinically promising means of predicting CKD. The framework is particularly appropriate for use in clinical decision-support systems where both accuracy and interpretability are required.

Authors

Institutions

Publication Details

Published
2026-11-21
DOI
https://doi.org/10.54117/wvmz6306
Primary Topic
Artificial Intelligence in Healthcare
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Hybrid Linear Discriminant Analysis and Artificial Neural Network Framework for High-Accuracy Prediction of Chronic Kidney Disease

Felix Ukpai Ogban, Otu Ofem Eteng, Okori David Etim
Artificial Intelligence in Healthcare
article

A Hybrid Linear Discriminant Analysis and Artificial Neural Network Framework for High-Accuracy Prediction of Chronic Kidney Disease

Felix Ukpai Ogban, Otu Ofem Eteng, Okori David Etim
article en

Abstract

Chronic kidney disease (CKD) is becoming an increasingly important global health issue and it is currently one of the major causes of death, predicted to become the fifth leading cause of years of life lost by 2040. It is challenging to detect it during its early stages as the disease is typically asymptomatic at this point, and nonrenal factors can interfere with the traditional diagnostic indicators. This paper describes a hybrid machine-learning approach that uses Linear Discriminant Analysis (LDA) for supervised dimensionality reduction and feature extraction and Bayesian-optimized Artificial Neural Network (ANN) for nonlinear classification. Using the publicly available UCI Chronic Kidney Disease dataset (consisting of 400 cases and 25 clinical and laboratory attributes), the method handles missing values, encodes the categorical variables, ranks the clinically relevant features, projects the data onto a maximally discriminative subspace, and classifies the patients using a regularized multilayer perceptron. The hybrid LDA–ANN model obtained an accuracy of 98.75% on the independent test set, with precision, recall, and F1-scores varying between 0.97 and 1.00, significantly beating logistic regression, decision trees, support vector machines, random forests, and standalone LDA (96.25%). The feature importance analysis showed that the selected predictors were clinically relevant features, comprising specific gravity, haemoglobin, albumin, diabetes status and hypertension. The findings show that the combination of LDA's ability to maximize class separability and the ANN's ability to model complex nonlinear interactions results in a robust, computationally efficient, and clinically promising means of predicting CKD. The framework is particularly appropriate for use in clinical decision-support systems where both accuracy and interpretability are required.

Vol. 1(1)
University of Cross River State (NG), University of Calabar (NG)
Artificial Intelligence in Healthcare
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
A Hybrid Linear Discriminant Analysis and Artificial Neural Network Framework for High-Accuracy Prediction of Chronic Kidney Disease — Felix Ukpai Ogban, Otu Ofem Eteng, et al. · (2026) | TGRS Research Map | TGRS