Mode of delivery is a major cause of maternal mortality. Mode of delivery encountered by women in labor room includes Normal delivery, Breech, Cesarean Section and Destructive Operation. In this study, Statistical pattern recognition using Nearest Neighbour approach was used to classify the mode of delivery using data obtained from a renowned Hospital in Sokoto, Nigeria, for 228 pregnant women. Condensed and k-Nearest Neighbourhood techniques were applied on the data using R Statistical Software. The result showed that, the misclassification percentage rate of Condensed is between 6.14% and 20.17%. In comparison, that of k-Nearest Neighbourhood was found to be between 4.39% to 6.25%. Therefore, it shows that k-Nearest Neighborhood can predict the mode of delivery of any pregnant woman if correct and accurate information about her antenatal care is available. The error rate of 0.0439 obtained from the techniques used was small, so the number of deviations from the actual class of the mode of delivery is minimal.
Key words: K- nearest neighbor (KNN), Maternal Delivery, Classification.
|
