Bootstrap aggregating multivariate adaptive regression spline for observational studies in diabetes cases

Bambang W. Otok*, Romy Y. Putra, Sutikno, Septia D.P. Yasmirullah

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

Background: Diabetes is a serious chronic disease that occurs either when the pancreas does not produce enough insulin, or when the body cannot effectively use the insulin it produces. Based on reports from the ministry of health, East Java Province has a Diabetes Melitus (DM) prevalence rate of 2.1%. This figure is greater than the prevalence rate in Indonesia, which is 1.5%. Materials and Methods: This study using Bootstrap Aggregating (Bagging) Multivariate Adaptive Regression Spline (MARS) method to analyze observational studies in diabetes cases. This study is aimed to analyze the factors that influence the complications type II diabetes and compare the level of accuracy between MARS and Bagging MARS. Results: The results showed that the probability of not occurring disease complications in patients is 0.708 and the occurrence of complications is 0.202. The variable that has the greatest influence is diabetes gymnastics. Type 2 DM patients with attending to diabetes gymnastics tend to not get disease complications as 1.857 times compared to patients with not attend to diabetes gymnastics. Conclusion: The accuracy of the classification between the MARS method and bagging MARS with 50, 100, 150, and 200 replications obtained the same results. This shows that bagging MARS cannot always improve accuracy.

Original languageEnglish
Pages (from-to)406-413
Number of pages8
JournalSystematic Reviews in Pharmacy
Volume11
Issue number8
DOIs
Publication statusPublished - 2020

Keywords

  • Accuracy
  • And Replication
  • Bagging MARS
  • Diabetes Melitus

Fingerprint

Dive into the research topics of 'Bootstrap aggregating multivariate adaptive regression spline for observational studies in diabetes cases'. Together they form a unique fingerprint.

Cite this