Temporal Analysis and Forecasting of Malaria Cases in Children Under 5 in Senegal's Tambacounda, Kolda, and Kédougou Regions: Using STL and Arima Models

Abstract

This study analyzes temporal trends and forecasts malaria cases in children under five years old in three regions of Senegal (Tambacounda, Kolda, and Kédougou), which are heavily affected by the disease. Using advanced statistical models, namely Seasonal and Trend decomposition using Loess (STL) and the Auto-Regressive Integrated Moving Average (ARIMA) model, the study aims to break down malaria time series into long-term trends, recurring seasonal patterns, and anomalies. The comparative performance of the STL and ARIMA models in predicting future cases is also assessed, offering insights to improve malaria control strategies.The results reveal a marked seasonal variation with a significant peak around 2018, followed by a stabilization of malaria cases, although regional differences were observed, with Kolda showing more pronounced fluctuations. The study highlights the importance of region-specific monitoring and control strategies and suggests that integrating additional data, such as climatic factors, could enhance future forecasts. This research provides a detailed understanding of malaria dynamics in these key regions, aiming to support more targeted and effective public health interventions.