Mathematical modeling to predict the anemia based on medical data

Abstract

Different diseases and diagnostic methods using various tests produced large amounts of complex medical data. Therefore, huge number of patient records in clinical centers, hospitals, and other health institutions have created the need for developed and accurate medical applications to help doctors. Since anemia is one of the most common health problems in recent era, the aim of this thesis is to predict anemia from a population through biomedical variables of individuals (the blood variables, age, and sex) and the anemia types using the currently produced mathematical models. This work is carried out using the dataset consisting of 539 subjects provided from blood laboratories. This thesis basically focuses on mathematical modeling to predict the anemia problem based on medical data. The main problems associated with medical diagnose involve the identification of highly accurate prediction models. For the first step, a mathematical method based on multiple linear regression (MLR) analysis has been applied to a reliable model that investigate if there exists a relation between the anemia and the biomedical variables and to provide the more realistic one. For the second step, a multiple nonlinear regression analysis has been used for a reliable model that research if there exists a mathematical relation between the observational variables and the anemia types. The parameter values produced are all seen to be the optimum values obtained from the multiple regression approaches, to provide the more realistic one. At the last step, optimum medical models based on biomedical variables are produced and an effective technique is used in investigating the optimum parameters of the models. To achieve this, the particle swarm optimization (PSO) algorithm has effectively been applied in predicting the parameters of the modxv els through the biomedical variables. Optimum values of the parameters produced from the PSO algorithm are used here to obtain more realistic models. The current models have been compared with the other ones and the results have been seen to be better. The models based on the variables and outcomes are expected to serve as a good indicator of disease diagnosis for health providers and planning treatment schedules for their patients. Thus, the study has been seen to be beneficial especially for those are interested in biomedical models arising in various fields of medical science, especially anemia.