Predicting Antenatal Care Follow Up Of Women During Pregnancy in Ethiopia Using Data Mining Technology
No Thumbnail Available
Date
2015-10-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
The purpose of this research is to show applicability of data mining techniques in EDHS 2011 data set and predict factors affecting antenatal care by using data mining techniques. In this research J48 algorithm, naïve Bayes algorithm and neural network were used in different percentage split and cross validation folds test environment. WEKA toolkit is used to build the models and performance of the classification algorithms on the dataset was measured by precision, recall, error rate and ROC curves.
Despite low utilization of health care services, there is a considerable variation across different demographic and socio-economic variables. The explanation of this diversity may be complex if we handled it traditionally. Hence there is a need to apply data mining technology to extract hidden patterns and regularities.
The experiments have been conducted following the Cios et al. data mining process model. After understanding the Problem domain and data, data preprocessing were done to handle missing values, to handle outliers, select relevant attributes and to smooth unbalanced EDHS 2011 dataset. The model developed by using pruned J48 with selected attribute on 80 percentage split showed highest classification accuracy. Naïve Bayes with selected attribute showed lowest time complexity and multilayer perceptron showed the highest time complexity. Despite low utilization of health care services, there is considerable variation across different demographic and socio-economic variables. If residence is rural having education and at least once a week frequency of listening radio then yes to antenatal care. If residence is urban from Amhara region with no education as well as no frequency of watching television and listening to radio, then no to antenatal care. Finally recommendation is given to develop a knowledge base that helps health center to use the knowledge extracted using data mining.
Description
Keywords
Predicting Antenatal Care Follow Up