A Comparison of Exponential Smoothening and ARIMA Modelling of the Infertility Rate among Women in Zaria Metropolis

International Journal of Research and Innovation in Applied Science (IJRIAS) | Volume VI, Issue X, October 2021|ISSN 2454-6194

A Comparison of Exponential Smoothening and ARIMA Modelling of the Infertility Rate among Women in Zaria Metropolis

 Oyenike Mary Olanrewaju, Mukhtar Umar Shitu*
Computer Science Department, Federal University Dutsin-Ma, Katsina, Nigeria
*Corresponding Author

Abstract. Infertility has recently become a socially destabilizing condition worldwide and it is considered as a serious problem in sub-Saharan Africa. This problem has caused conflict and marital disharmony among couples due to stigma and other social discriminations. Although, infertility affects both men and women, the social and psychological trauma affects more women than men. The clinical pattern of infertility varies in different parts of the world, and Nigeria has recorded more cases in Africa. Diagnosis and treatment of infertility is often viewed as a low priority in national development policies and health strategies of the public hospitals. The lack of funding, effective allocation of available health resources and standard infrastructures by the government as well as by international sponsors to the public health sector remain a challenge in most of our societies today. This research compares the optimization of two time-series approaches namely, an autoregressive integrated moving average (ARIMA) and Holt-Winters exponential smoothing in predicting infertility. The research evaluates the forecasting performance of the models using the post forecast summary statistics which includes; mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE). The result indicates that exponential smoothing out-performed the ARIMA model with a lower MAE (4.16), RSME (5.27) and MAPE (16.15). Although both models performed well, the selected model gave the optimal accurate prediction compared to the other model. The validity of the model was further checked by comparing the fitted values with the actual values.

Key words: Infertility, Autoregressive Integrated Moving Average (ARIMA), Holt-Winters Exponential Smoothing, Trends, mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE).

I. INTRODUCTION

Infertility is a reproductive system disease that affects both male and female of childbearing age. It is defined as a medical condition characterized by the inability to conceive after 12 months of unprotected sexual intercourse [1]. Globally, it is estimated that 48.5 million couples experience infertility every year [2]. In Africa, because of the level of literacy, infertility is often linked with different superstition held by some people [3].

Read More