A Comparative Study of Multiple Linear Regression-Clustering-SVM and Fuzzy Linear Regression-Symmetric Parameter Clustering-SVM Hybrid Models in Predicting Colorectal Cancer

Nur Ain  Ebas; Muhammad Ammar  Shafi; Mohd Saifullah  Rusiman; Toh  Yoke Teng; Zeety  Md Yusof; Nurain Izzati  Mohd Yassin; Banan Badeel  Abdal

Authors

Nur Ain Ebas Department of Civil Engineering, Faculty of Civil Engineering and Built Environment, Universiti Tun Hussein Onn Malaysia
Muhammad Ammar Shafi Department of Technology Management and Business, Universiti Tun Hussein Onn Malaysia
Mohd Saifullah Rusiman Department of Mathematics and Statistics, Faculty of Applied Sciences and Technology, Universiti Tun Hussein Onn Malaysia
Toh Yoke Teng Department of Civil Engineering, Faculty of Civil Engineering and Built Environment, Universiti Tun Hussein Onn Malaysia
Zeety Md Yusof Department of Civil Engineering, Faculty of Civil Engineering and Built Environment, Universiti Tun Hussein Onn Malaysia
Nurain Izzati Mohd Yassin Department of Civil Engineering, School of Engineering and Computing, MILA University, Nilai, Negeri Sembilan, Malaysia
Banan Badeel Abdal College of Administration and Economics, University of Duhok, Iraq

Keywords:

colorectal cancer, fuzzy linear regression, multiple linear regression, hybrid model, error metrics

Abstract

Colorectal cancer (CRC) remains a leading cause of mortality worldwide, with early detection being crucial for improving patient outcomes. In order to predict the colorectal cancer, this study compares two hybrid machine learning models, which are Multiple Linear Regression Clustering with Support Vector Machine (MLRCSVM) and Fuzzy Linear Regression with Symmetric Parameter Clustering with Support Vector Machine (FLRWSPCSVM). Secondary data was obtained from a general hospital in Kuala Lumpur. It includes 180 colon cancer patients as respondents, with data collected and recorded by nurses using cluster sampling. The size of the tumor is the dependent variable, while colorectal cancer symptoms and factor are the independent variables. Mean square error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE) are used to evaluate the performance of these models. The results indicate that FLRWSPCSVM outperforms MLRCSVM in terms of accuracy and robustness in handling uncertain or noisy data, highlighting its potential as a powerful tool for early colorectal cancer diagnosis.