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

Feature selection and classification model construction on type 2 diabetic patients' data.

Yue Huang1, Paul McCullagh, Norman Black

  • 1Department of Computing, Faculty of Engineering, Imperial College London, South Kensington, London SW7 2AZ, UK. y.huang@imperial.ac.uk

Artificial Intelligence in Medicine
|August 21, 2007
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

FUT8 Is a Critical Driver of Prostate Tumour Growth and Can Be Targeted Using Fucosylation Inhibitors.

Cancer medicine·2025
Same author

Data Mining and Fusion Framework for In-Home Monitoring Applications.

Sensors (Basel, Switzerland)·2023
Same author

ST6GAL1-mediated aberrant sialylation promotes prostate cancer progression.

The Journal of pathology·2023
Same author

Upregulation of GALNT7 in prostate cancer modifies O-glycosylation and promotes tumour growth.

Oncogene·2023
Same author

Anaplastic lymphoma kinase (ALK)-positive large B-cell lymphoma: a rare cause of ileal intussusception.

BMJ case reports·2022
Same author

Challenging the 50-50 rule for the basal-bolus insulin ratio in patients with type 2 diabetes who maintain stable glycaemic control.

Diabetes, obesity & metabolism·2022

This study identified key factors like age, diagnosis duration, and insulin treatment influencing diabetes control. Data mining techniques achieved 95% accuracy in predicting patient outcomes, aiding public health awareness and prevention efforts.

Area of Science:

  • Medical Informatics
  • Public Health
  • Data Science

Background:

  • Diabetes affects millions globally, posing significant public health and economic challenges.
  • Effective diabetes management requires understanding key influencing factors.
  • Advancements in patient management systems generate large datasets for analysis.

Purpose of the Study:

  • To identify significant factors impacting diabetes control using feature selection.
  • To develop classification models for predicting poor diabetes control status.
  • To leverage data mining for knowledge discovery in diabetes management.

Main Methods:

  • Applied feature selection (FSSMC, an optimization of ReliefF) to identify important attributes.
  • Utilized data mining techniques on patient data collected from 2000-2004.

Related Experiment Videos

  • Employed three classification techniques (Naïve Bayes, IB1, C4.5) for prediction.

    • Main Results:

    • Identified 'age', 'diagnosis duration', 'insulin treatment', 'random blood glucose', and 'diet treatment' as key factors.
    • Achieved a predictive accuracy of 95% and sensitivity of 98% with reduced features.
    • Highlighted the influence of 'type of care', 'home monitoring', and 'smoking' on outcomes.

    Conclusions:

    • Factors like age and diagnosis duration are physician-independent, while insulin, diet, and lifestyle (BMI, smoking) are controllable.
    • Emphasizes the need for public health education and patient empowerment.
    • Data mining serves as a valuable exploratory tool for understanding complex health data and informing clinical practice.