TY - GEN
T1 - Predictive Modeling of Cirrhosis Patient Survival using Random Forest Algorithm
AU - Thatha, Naganikhila
AU - Seenivasan, E.
AU - Bharath, A.
AU - Arunsankar, G.
AU - Archana, N.
AU - Sathishkumar, V. E.
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - Cirrhosis is a chronic hepatic disorder that markedly elevates mortality risk if not addressed with prompt and precise prognostic instruments. This research introduces a machine learning methodology employing the Random Forest algorithm to forecast survival outcomes in patients with cirrhosis. Employing a publicly accessible dataset from the Mayo Clinic, comprising of 418 patient records and 17 clinical parameters, a prediction model is constructed and validated to categorize patients into three classifications: death, censored (alive), and censored due to liver transplantation. The dataset undergoes preprocessing, which includes addressing missing values, encoding categorical variables, and normalizing numerical data. The proposed Random Forest model attains a classification accuracy of 99.04% and an AUC of 0.98, surpassing baseline classifiers. Feature significance analysis identified bilirubin, albumin, prothrombin time, and ascites as the primary predictors. The results indicate that Random Forest presents a reliable and interpretable method for predicting survival in cirrhosis patients, hence offering significant decision assistance for physicians.
AB - Cirrhosis is a chronic hepatic disorder that markedly elevates mortality risk if not addressed with prompt and precise prognostic instruments. This research introduces a machine learning methodology employing the Random Forest algorithm to forecast survival outcomes in patients with cirrhosis. Employing a publicly accessible dataset from the Mayo Clinic, comprising of 418 patient records and 17 clinical parameters, a prediction model is constructed and validated to categorize patients into three classifications: death, censored (alive), and censored due to liver transplantation. The dataset undergoes preprocessing, which includes addressing missing values, encoding categorical variables, and normalizing numerical data. The proposed Random Forest model attains a classification accuracy of 99.04% and an AUC of 0.98, surpassing baseline classifiers. Feature significance analysis identified bilirubin, albumin, prothrombin time, and ascites as the primary predictors. The results indicate that Random Forest presents a reliable and interpretable method for predicting survival in cirrhosis patients, hence offering significant decision assistance for physicians.
KW - Cirrhosis
KW - Clinical Data Analysis
KW - Liver Disease
KW - Machine Learning
KW - Medical Decision Support
KW - Prognostic Modeling
KW - Random Forest
KW - Survival Prediction
UR - https://www.scopus.com/pages/publications/105035594135
U2 - 10.1109/IC2NC67409.2025.11376423
DO - 10.1109/IC2NC67409.2025.11376423
M3 - Conference contribution
AN - SCOPUS:105035594135
T3 - International Conference on NexGen Networks and Cybernetics, IC2NC 2025 - Proceedings
SP - 952
EP - 957
BT - International Conference on NexGen Networks and Cybernetics, IC2NC 2025 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2025 International Conference on NexGen Networks and Cybernetics, IC2NC 2025
Y2 - 1 December 2025 through 3 December 2025
ER -