GENERIC PREDICTION OF MALARIA TREATMENT OUTCOMES USING BIG DATA ANALYTICS (BDA)

GENERIC PREDICTION OF MALARIA TREATMENT OUTCOMES USING BIG DATA ANALYTICS (BDA)

Author by Dr. Oluwaseun Ebiesuwa

Journal/Publisher: Proceedings Of 26th Nigeria Computer Society National Conference

Volume/Edition: 27

Language: English

Pages: 75 - 83

Abstract

Research has shown that the synergy between big data analytics and healthcare enhances the quality of care on a patient by patient basis as well as a massive reduction in expenditure made on account of healthcare-related problems.  This study addresses malaria disease, a perennial problem that has plagued the vast majority of people in Africa, especially as it pertains to ascertaining the best drug combination and progression that should be followed by physicians in the treatment so as to provide get the best outcomes and provide wholesome healthcare for those suffering from malaria. The study implemented big data analytics presented in the National Prediction Framework, in order to predict best effective anti-malarial drug combination to countries plagued with the disease. Large malaria data was sourced from the Institute of Child Health, University College Hospital, Ibadan – Nigeria. The dataset was imported into WEKA (Waikato Environment for Knowledge Analysis) and pre-processed; the relevant attributes in the dataset used for this study are 22 in number. The Hadoop MapReduce framework was employed for this research because of the enormity of the data captured in terms of volume. The association data mining technique was applied in order to relate patients’ symptoms with the medication; this association between the patients’ symptoms and the medication was done using the APRIORI algorithm for mining the data and for generating the best ten rules during the rule selection phase of mining. The rule extraction phase succeeded the rule selection phase in which the most pertinent rules needed for prediction of malaria treatment outcomes were extracted and subsequently, inferences were generated as regards the better progression of drug use that should be adopted by a patient infected with malaria. Using WEKA software, the best ten rules were generated and results showed that the adopted rules give the best progression to be followed in the treatment of malaria using combination treatment approach. Consequently, findings in this study lends credence to the fact that existing big data mining techniques can generate reliable results which can be very instructive for healthcare practitioners in general to help salvage third world countries particularly in Sub-Saharan Africa from the malaria menace that claims many innocent lives daily.


Other Co-Authors