From Preprocessing to Prediction: An Analytical Study on Diabetes Data

Title

From Preprocessing to Prediction: An Analytical Study on Diabetes Data

Creator

Siju, Marbin; Bose, Shreya; Vijayalakshmi, A.

Description

Early detection of diabetes is crucial for improving a patients long-term health. In this chapter, we study diabetes and diabetes-related factors. We also delve into various imputation techniques used to address missing data. Missing data is generally a very critical issue in healthcare analytics, as a limited history of clinical records often leads to biased analysis and suboptimal model representation. This chapter gives a detailed literature review of data imputation methods. In this chapter, we have done two case studies. In the first case study, mean, median, and mode imputation techniques are applied to artificially created missing values to examine their effect on the structure and distribution of the data. The second case study captures a prediction model for a diabetes diagnosis using the same dataset. Here, a random forest prediction model is created to predict the possible presence of diabetes. An accuracy of 97.07% is achieved on the test data, which shows that diabetes can be predicted by considering other dependable variables. 2026 selection and editorial matter, Syed Nisar Hussain Bukhari; individual chapters, the contributors.

Source

Artificial Intelligence for Predictive Healthcare: Towards Personalized Treatment and Disease Prevention;pp.33-62

Date

01-01-2026

Publisher

CRC Press

Coverage

Siju M., Department of Statistics and Data Science, CHRIST University, Bangalore, India; Bose S., Department of Statistics and Data Science, CHRIST University, Bangalore, India; Vijayalakshmi A., Department of Statistics and Data Science, CHRIST University, Bangalore, India

Rights

Restricted Access; Hardcopy may be available in the library

Relation

ISBN: 978-104058017-2; 978-100367454-2;

Format

online

Language

English

Type

Book chapter

Identifier

https://doi.org/10.1201/9781003674542-3

https://www.scopus.com/pages/publications/105034236743?origin=resultslist