Related Experiment Video
Updated: Jun 16, 2026
13:44
Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns
Published on: August 30, 2013
Breast-cancer identification using HMM-fuzzy approach
Md Rafiul Hassan1, M Maruf Hossain, Rezaul Karim Begg
1Department of Computer Science and Software Engineering, The University of Melbourne, Victoria 3010, Australia. mrhassan@csse.unimelb.edu.au
Computers in Biology and Medicine
|February 16, 2010
Summary
This study introduces a novel hybrid Hidden Markov Model (HMM)-fuzzy approach for breast lesion classification. The method effectively distinguishes benign from malignant tumors using optimized fuzzy rules and selected features.
Area of Science:
- Biomedical Engineering
- Machine Learning
- Medical Informatics
Background:
- Accurate classification of breast lesions is crucial for effective cancer treatment.
- Existing computational methods face challenges in achieving high diagnostic accuracy.
Purpose of the Study:
- To develop and evaluate an ensemble feature selection and classification technique for breast lesion diagnosis.
- To enhance classification performance using a hybrid Hidden Markov Model (HMM)-fuzzy approach.
Main Methods:
- Feature selection based on Area Under the ROC Curves (AUC).
- Classification using a hybrid Hidden Markov Model (HMM)-fuzzy approach.
- Optimization of fuzzy rules using gradient descent algorithms.
Main Results:
- The developed model effectively classifies benign and malignant breast lesions.
- Achieved high classification accuracy using only two optimized fuzzy rules.
- Demonstrated superior performance compared to other computational tools on the Wisconsin breast cancer dataset.
Conclusions:
- The proposed HMM-fuzzy approach combined with feature selection offers a powerful tool for breast lesion classification.
- The model's efficiency and accuracy highlight its potential for clinical application.
- Further research can explore its applicability to other medical diagnostic tasks.
