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

Appropriate medical data categorization for data mining classification techniques.

Shang-Chih Liao1, I-Nong Lee

  • 1Department of Internal Medicine, Chang Gung Memorial Hospital, Kaohsiung, Taiwan.

Medical Informatics and the Internet in Medicine
|January 2, 2003
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

Update on the Application of Ultrasonography in Understanding Autosomal Dominant Polycystic Kidney Disease.

Journal of medical ultrasound·2024
Same author

Clinical effectiveness of molnupiravir in patients with COVID-19 undergoing haemodialysis.

International journal of antimicrobial agents·2023
Same author

Reference ranges for ultrasonographic renal dimensions as functions of age and body indices: A retrospective observational study in Taiwan.

PloS one·2019
Same author

Longitudinal analysis of cardiac structure and function in incident-automated peritoneal dialysis: comparison between icodextrin solution and glucose-based solution.

BMC nephrology·2018
Same author

Incidence and outcomes of encapsulating peritoneal sclerosis (EPS) and factors associated with severe EPS.

PloS one·2018
Same author

Corrigendum: An association between time-varying serum alkaline phosphatase concentrations and mortality rate in patients undergoing peritoneal dialysis: a five-year cohort study.

Scientific reports·2017
Same journal

Patients' perceptions of Internet usage and their opportunity to obtain health information.

Medical informatics and the Internet in medicine·2007
Same journal

Assisting the transition from hospital to home for children with major congenital heart disease by telemedicine: a feasibility study and initial results.

Medical informatics and the Internet in medicine·2007
Same journal

Digital pens and pain diaries in palliative home health care: professional caregivers' experiences.

Medical informatics and the Internet in medicine·2007
Same journal

Readability and cultural sensitivity of web-based patient decision aids for cancer screening and treatment: a systematic review.

Medical informatics and the Internet in medicine·2007
Same journal

ICT-based health information services for elderly people: past experiences, current trends, and future strategies.

Medical informatics and the Internet in medicine·2007
Same journal

Obtrusiveness of information-based assistive technologies as perceived by older adults in residential care facilities: a secondary analysis.

Medical informatics and the Internet in medicine·2007
See all related articles

Data categorization enhances data mining (DM) for medical data. The minimum description length principle (MDLPC) method improved performance across most DM classification techniques, making results more trustworthy.

Area of Science:

  • Data mining
  • Medical informatics
  • Machine learning

Background:

  • Data mining methods vary in data format requirements (normalized, categorized, or multiple scales).
  • Each data mining technique has unique theoretical underpinnings, influencing results based on data format.
  • Categorical medical data can aid decision-making and knowledge extrapolation.

Purpose of the Study:

  • To identify optimal data formats for diverse data mining classification techniques.
  • To enhance the efficiency and trustworthiness of data mining results in medical applications.
  • To explore the impact of data categorization on classification accuracy.

Main Methods:

  • Applied three mathematical data categorization methods: Fusinter, Minimum Description Length Principle (MDLPC), and Chi-merge.

Related Experiment Videos

  • Utilized five data mining classification techniques: discriminant analysis, Neural Networks, Decision Trees, Genetic Supervised Clustering, and Bayesian classification (Probability Neural Networks; PNN).
  • Tested methods on a heart disease database with continuous, binary, nominal, and ordinal data types.

    • Main Results:

    • Data categorized using the MDLPC method demonstrated superior performance across most data mining classification techniques compared to original or normalized data.
    • Categorical data generally proved beneficial for most data mining classification techniques, particularly for classifying disease and non-disease groups.
    • MDLPC-categorized data facilitated easier extraction of medical knowledge.

    Conclusions:

    • Data categorization, especially using the MDLPC method, is crucial for optimizing data mining classification in medical contexts.
    • Categorical data formats enhance the usability and interpretability of data mining results for medical knowledge discovery.
    • The study provides guidance on selecting appropriate data formats for different data mining techniques to achieve reliable outcomes.