Non-Communicable Diseases in Pregnant Women Based on A Behavioral Approach

Abstract

Background: Early detection of pregnant women as a prevention of the risk of non-communicable diseases can be done with routine health checks. The general aim of this research is to produce a program from an artificial intelligence system to detect non-communicable diseases early and provide WhatsApp-based recommendations to pregnant women.
 Method: The implementation of this research began by creating a dataset obtained from the Medical Records Installation, namely data on pregnant women for 3 years from 2019 to mid-2023. Then the data obtained was coded, processed, and classified according to research needs, resulting in 9,289 data. The data is entered into machine learning to be processed by the machine to determine the mean risk factors, which will then produce prediction data. The first stage in data processing required is a machine learning application which will be used to process big data into predictions.
 Result: In this research, the application used is Google Collab, which is a default application from Google and can be used with various devices. In this study, the dataset used by researchers is a dataset that predicts heart disease, hypertension, preeclampsia, and eclampsia and recommendations for pregnant women that provide good performance on each accuracy test. After the first process of data sharing, the training data is 90% and the 10% data is called testing data.
 Conclusion: The data obtained from pregnant women is then processed to obtain quality data by applying data cleaning using a scaler, namely data whose attribute values ​​will be empty so that the data becomes more accurate. A pregnant woman dataset of 9289 records with complete attributes of 9289 records will be used in the experimental process.