Constructing A Predictive Model For Occurrence Of Tuberculosis The Case Of Menelik Ii Hospital And St. Peters Tb Specialized Hospital.

ABSTRACT rnBackground: Tuberculosis is a disease of poverty affecting mostly young adults in their mostrnproductive years. In Ethiopia, TB is a disease of major public health problem. Earlyrnidentification and isolation of TB cases is critical to prevent further transmission, morbidity andrnmortality caused by TB. Data mining has a potential to indentify hidden knowledge from hugerndatasets. It is possible to use data mining algorithms for analysis and predicting the TB status ofrnpatients. rnObjective: The goal of this research was to apply data mining techniques for predicting the TBrnstatus of patients. Specifically, identify the determinant attributes of TB status of patients, buildrnbest prediction model and finally develop a prototype graphical user interface. rnMethodology: A hybrid data mining process model that involved six steps is followed. Thisrnstudy considers a total of 10,031 records from Menelik II and St. Peters TB specialized hospitalsrnpatients’ data and 15 attributes for predicting the TB status. Descriptive data analysis,rnvisualization and statistical summary were implemented to gain understanding of the data.rnHandling of missing values and data transformation were done to prepare the dataset forrnexperimentation. The mining algorithms used are decision tree, naïve bayes, support vectorrnmachine and artificial neural network. To evaluate the models performance 10-fold crossrnvalidation and confusion matrix are used. rnResults: The result of the experiments with all and selected attributes showed that performancernof J48, Sequential minimal optimization and Multilayer perceptron were better with all attributesrnthan best selected attributes, whereas naïve bayes classifier performance increased with selectedrnattributes than all attributes. The results of the experiments show the performance of miningrnalgorithms decreases as the amount of training increases. rnThe best selected model to predict the TB status of patients in this study was generated by J48rndecision tree with all attributes. The accuracy of this model is 95.24%. Graphical user interfacernprototype was designed using the ten rules from J48 decision tree. rnConclusion: The results achieved from this research indicate that data mining is useful inrnbringing relevant information from large and complex patients’ dataset, and we can use thisrninformation for predicting TB status and decision making. The most important attributes thatrndetermine the TB status of the patients are shortness of breath, chest pain, cough, weight loss,rnloss of appetite, night sweats and HIV test results.

Get Full Work

Report copyright infringement or plagiarism