Temporal prediction of malaria in Apartadó through a probabilistic random walk

Abstract

Introduction: the analysis of the behavior of malaria dynamics has helped to develop causal methods which have very accurately predicted the number of people infected with this disease annually.

Objective: to predict the monthly behavior of malaria in the city of Apartadó, Colombia, using the probabilistic random walk.

Method: the number of cases of malaria registered in Apartadó and reported monthly by SIVIGILA were obtained, and their analogy with the random walk was evaluated. Then, the nonequiprobability of the phenomenon obtaining monthly predictions using a second-class equation was determined.

Results: the analogy between the random walk and the monthly behavior of malaria in Apartadó was proven, and therefore the method was used to predict the number of cases in consecutive months in 2018, achieving a predictive accuracy ranging from 84.4-100 %.

Conclusions: the behavior of malaria in the city of Apartadó has mathematical sequences that allow for its spatial-temporal prediction using the probabilistic random walk, allowing it to be used to guide epidemiological surveillance activities and observe the effectiveness of interventions on public health.