Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

An approach for generating fuzzy rules from decision trees.

Amir R Razavi1, Mikael Nyström, Marian S Stachowicz

  • 1Dept. of Biomedical Engineering, Div. of Medical Informatics, Linköping University, Sweden. amirreza.razavi@imt.liu.se

Studies in Health Technology and Informatics
|November 17, 2006
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Does SNOMED CT post-coordination scale?

Studies in health technology and informatics·2014
Same author

User-directed coordination in SNOMED CT.

Studies in health technology and informatics·2013
Same author

Applying representational state transfer (REST) architecture to archetype-based electronic health record systems.

BMC medical informatics and decision making·2013
Same author

Coding of procedures documented by general practitioners in Swedish primary care-an explorative study using two procedure coding systems.

BMC family practice·2012
Same author

A decision support model for cost-effectiveness of radical prostatectomy in localized prostate cancer.

Scandinavian journal of urology and nephrology·2011
Same author

The nature of unintended effects of health information systems concerning patient safety: a systematic review with thematic synthesis.

Studies in health technology and informatics·2010

This study developed a fuzzy model from crisp rules for breast cancer recurrence prediction. The fuzzy approach showed similar accuracy to crisp rules, offering a more robust and understandable clinical tool.

Area of Science:

  • Oncology
  • Machine Learning
  • Computational Biology

Background:

  • Accurate identification of high-risk breast cancer patients is crucial for effective clinical management and patient outcomes.
  • Tumor characteristics like size are key indicators but can benefit from nuanced analysis.

Purpose of the Study:

  • To develop and evaluate a fuzzy rule-based model for predicting breast cancer recurrence.
  • To compare the predictive performance of a fuzzy model against traditional crisp rules derived from Decision Tree Induction (DTI).

Main Methods:

  • Decision Tree Induction (DTI) was applied to a dataset of 3949 female breast cancer patients to generate crisp If-Then rules.
  • Crisp rules were converted into fuzzy rules by assigning membership functions to variables, creating a fuzzy mathematical model.

Related Experiment Videos

  • The fuzzy model's predictions were compared against crisp rule predictions using Area Under the ROC Curve (AUC) on 100 randomly selected cases.

    • Main Results:

    • No significant difference in predictive accuracy for breast cancer recurrence was observed between the fuzzy rule-based model and the original crisp rules.
    • The fuzzy model demonstrated robustness to noise and improved comprehensibility for clinicians.

    Conclusions:

    • Fuzzy rule-based systems derived from DTI offer a viable alternative for breast cancer recurrence prediction.
    • Soft discretization of variables using fuzzy logic enhances model interpretability and robustness, aiding clinical decision-making.