Projecting Teenage Pregnancy and Motherhood and Fertility Rates: An Analytical and Machine Learning Study of Maharashtra, India

Abstract

Adolescent pregnancy and motherhood is a huge problem with building physical, mental, and economic implications for teenage mothers’ futures and their babies as well.  This paper is divided into 2 halves, fertility trends and teenage pregnancy and motherhood, both in Maharashtra. All the data tables for this paper are taken from the 5 phase NFHS surveys. A dataset from each of the reports for the 5 phases were integrated for the first part whereas the most recent phase – the fifth phase (NFHS-5) – was taken for the teenage pregnancy and motherhood part. EDA, Exploratory Data Analysis was used for the analysis and visualisation of this paper. The variable correlated in this study are residency, age, marital status, education, TFR, CBR, religion, and caste and tribe to find the answers to the 7 questions. The findings of this paper suggest that higher rates of pregnancy are related to rural areas of residence, lower income levels, lesser education, and being a part of a scheduled tribe or being a Hindu. Furthermore, machine learning modelling, Linear Regression algorithm, was employed on collected data. The performance of the built model was calculated using five metrics. The lower mean absolute error and root mean squared error values indicated that the developed regression model is more accurate.