Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM67452
Volume:
10 |
9 |
Sep 2026
Pages:
1-7
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
computer Science
Institute of Technology and Management, Lucknow
Liver disease screening models are often evaluated as classifier comparisons rather than as clinical prediction systems. The supplied literature shows two related problems. A 2020 study evaluated six algorithms on 583 records from the Indian Liver Patient Dataset using ten-fold cross-validation and reported accuracy, kappa, mean absolute error, and execution time, but did not report external validation, calibration, confidence intervals, or decision utility. A 2019 corrigendum to another liver classification study corrected the placement of false-positive and false-negative counts and consequently revised multiple metrics. A clinical review also describes substantial etiologic and geographic heterogeneity in liver disease. This paper proposes a reproducible protocol for developing and validating a routine-data screening model across sites and time. The design uses patient-level separation, nested cross-validation, fold-contained preprocessing and feature selection, class-sensitive learning, probability calibration, temporal testing, external validation, subgroup analysis, and an automated confusion-matrix audit. Logistic regression, elastic-net logistic regression, random forest, and gradient-boosted trees are compared using discrimination, calibration, threshold performance, and decision-curve net benefit. No patient-level data or experimental outputs were supplied; therefore, numerical results are not reported. The contribution is a validation and reporting framework that directly addresses the limitations supported by the reviewed documents without claiming unobserved performance.
KEYWORDS
The supplied literature supports a clear methodological gap. A small single-dataset classifier comparison cannot establish that liver-disease predictions generalize across etiologies, time, or institutions, and the published correction of FP/FN placement demonstrates the need for auditable metric generation. This paper proposes a multicenter protocol that combines fold-contained preprocessing, nested model development, temporal and external validation, calibration, decision analysis, subgroup evaluation, explainability, and automatic confusion-matrix checks. The next step is to secure approvals, finalize the adjudication manual, and execute the locked analysis. The completed study should publish all prespecified tables and figures regardless of whether complex models outperform logistic regression. If external calibration or net benefit is inadequate, future work should focus on outcome definition, predictor harmonization, or local recalibration before considering a more complex algorithm. A subsequent prospective study should measure whether use of the model changes referral decisions or patient outcomes.
[1] R. Williams, "Global challenges in liver disease," Hepatology, vol. 44, no. 3, pp. 521-526, Sept. 2006, doi: 10.1002/hep.21347. [2] J. Singh, S. Bagga, and R. Kaur, "Software-based prediction of liver disease with feature selection and classification techniques," Procedia Computer Science, vol. 167, pp. 1970-1980, 2020, doi: 10.1016/j.procs.2020.03.226. [3] M. Abdar, M. Zomorodi-Moghadam, R. Das, and I.-H. Ting, "Corrigendum to Performance Analysis of Classification Algorithms on early detection of Liver disease [Expert Systems with Applications 67 (2017) 239-251]," Expert Systems with Applications, vol. 125, pp. 442-443, 2019, doi: 10.1016/j.eswa.2019.02.029. [4] G. S. Collins, J. B. Reitsma, D. G. Altman, and K. G. M. Moons, "Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis: The TRIPOD statement," Annals of Internal Medicine, vol. 162, no. 1, pp. 55-63, 2015, doi: 10.7326/M14-0697. [5] R. F. Wolff et al., "PROBAST: A tool to assess the risk of bias and applicability of prediction model studies," Annals of Internal Medicine, vol. 170, no. 1, pp. 51-58, 2019, doi: 10.7326/M18-1376. [6] E. R. DeLong, D. M. DeLong, and D. L. Clarke-Pearson, "Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach," Biometrics, vol. 44, no. 3, pp. 837-845, 1988, doi: 10.2307/2531595. [7] A. J. Vickers and E. B. Elkin, "Decision curve analysis: A novel method for evaluating prediction models," Medical Decision Making, vol. 26, no. 6, pp. 565-574, 2006, doi: 10.1177/0272989X06295361. [8] S. M. Lundberg and S.-I. Lee, "A unified approach to interpreting model predictions," in Advances in Neural Information Processing Systems 30, 2017, pp. 4765-4774.+
Publication Date
9/24/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
2
Total views since publication
PDF Downloads
3
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...