Predicting the Risk of Gestational Diabetes Mellitus using Data Mining Techniques
Abstract
Medical diagnosis can be done very effectively by applying knowledge discovery in medical databases. Data Mining is an effective technique used to extract knowledge from databases and also helps to generate unknown and hidden patterns from the information stored in the databases. Healthcare dataset used in this research is gestational women data taken during first trimester of the gestational period. Application of data mining algorithms in the gestational women dataset for predicting the risk of gestational diabetes is a newapproach in the research field. In order to perform data preprocessing for the gestational women dataset used in this research paper, equal width binning interval approach is used to discretize the continuous valued attributes. The desired width of the interval in the database is fixed after getting opinion from the medical experts. The discretized input values are given as input to the generated model. In this research, the process is done at two phases. In phase 1, the numerical attributes are converted into categorical values using discretization techniques and in phase 2, Apriori algorithm is applied to the database that generates association rules which are useful to identify general association in the gestational diabetes dataset. These generated association rules show the relationship among the measured attributes and also indicate the risk level of gestational diabetes.
Authors: Srideivanai Nagarajan, P. Ramasubramanian, S. Hariharan
Published in: World Congress on Sustainable Technologies (WCST-2021)
- Date of Conference: 7-9 December 2021
- DOI: 10.20533/WCST.2021.0012
- ISBN: 978-1-913572-41-9
- Conference Location: Virtual (London, UK)