AI-based risk prediction and dietary guidance for pregnant women using random forest models · Electronic Theses and Dissertations

Author

Mudamburi, Linnon

Title

AI-based risk prediction and dietary guidance for pregnant women using random forest models

Abstract

Pregnancy is a crucial period during which women require specific nutrients for fetal growth and development. To minimize complications, pregnant women with gestational diabetes or preeclampsia need specific dietary recommendations. Gestational diabetes and preeclampsia are both medical conditions that can affect pregnant women, resulting in adverse pregnancy outcomes, such as preterm birth, low birth weight, and developmental disorders. Artificial intelligence (AI) has become a promising technique for creating personalized dietary models that can offer individualized recommendations based on personal health data in recent years.
In this research, the researcher developed an AI-powered dietary model that provides nutritional management assistance to pregnant women with diabetes or preeclampsia, as well as those without, to promote healthy eating and prevent nutrient deficiencies. This research used a dataset that includes medical records and dietary data gathered from pregnant women. The researcher trained the AI-powered dietary model on a dataset of nutritional requirements during pregnancy to understand relationships between different variables and dietary intake.
The dataset was pre-processed to clean up, and the researcher arranged the data before training the model. To ensure that all variables were on the same scale, the researcher eliminated missing or pointless data points, standardized the data format, and normalized the data. The AI-powered dietary model was then created using machine learning techniques such as decision trees and random forest algorithms after the dataset had been pre-processed. The model was then put to the test to evaluate its accuracy.
Before the model was used to provide nutritional advice to pregnant mothers with or without diabetes or preeclampsia, the researcher improved the model's performance. The researcher highlighted the limitations that were associated with creating such a model. Based on the limitations, the researcher recommended that future researchers expand data sets and enhance the comprehensiveness of data to ensure that the research phenomenon is largely covered in terms of its scope.
Without the contributions of these individuals and organizations, this research project would not have been possible. I am deeply grateful for their support and look forward to continuing my work with their guidance and assistance.

Date

June 2025

Publisher

BUSE

Keywords

Pregnancy women

Supervisor

Mr. C. Zano