Related Experiment Video
Updated: Feb 27, 2026
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
8.2K
An enhanced heart disease prediction model based on linear Diophantine fuzzy-integrated supervised machine learning.
Jeevitha Kannan1,2, Vimala Jayakumar3, Nasreen Kausar4,5
1Center of Computational Biology, SRM Institute of Science and Technology, Ramapuram, Chennai, India.
Health Information Science and Systems
|February 26, 2026
Summary
This study introduces a novel hybrid model integrating the Linguistic Decision Fuzzy Sets (LDFS) framework with machine learning (ML) for improved heart disease diagnosis. The LDFS-ML approach effectively handles data ambiguity, enhancing diagnostic accuracy and interpretability.
Area of Science:
- Medical Informatics
- Artificial Intelligence in Medicine
- Fuzzy Logic Systems
Background:
- Conventional machine learning (ML) methods struggle with uncertainty and vagueness inherent in medical diagnoses.
- Existing ML pipelines require complex feature engineering for handling diverse data types (categorical and numerical).
Purpose of the Study:
- To develop and evaluate a hybrid model integrating Linguistic Decision Fuzzy Sets (LDFS) with ML algorithms for enhanced heart disease diagnosis.
- To address the limitations of conventional ML in handling ambiguous medical data and diverse feature types.
Main Methods:
- A novel hybrid LDFS-ML model was proposed, accommodating both categorical and numerical features using membership functions.
- Several ML algorithms (logistic regression, decision tree, SVM, XGBoost) were evaluated on both crisp and LDF-based datasets.
- Comparative analysis was performed to assess the performance of the LDFS-ML model against conventional ML approaches.
Main Results:
- The proposed LDFS-ML model consistently outperformed conventional ML algorithms in classification tasks.
- Significant improvements were observed in performance metrics on LDF-datasets, including a 0.97% increase in accuracy, 0.95% in precision, 0.99% in recall, and 0.97% in F1 score (XGBoost).
- The LDFS framework demonstrated effective handling of data ambiguity and simplified feature type management.
Conclusions:
- The integration of LDFS with ML offers a promising new direction for medical diagnosis, particularly in managing ambiguity.
- The LDFS-ML hybrid model enhances decision-making processes by providing improved interpretability and performance.
- This approach overcomes key limitations of conventional ML in medical applications, paving the way for more robust diagnostic tools.
