Decision Tree based Smart System for Pregnant Women Diagnosis

Abstract

one of the reasons of women and newborns death in developing countries is the complications during pregnancy and childbirth. An early diagnosis of pregnancy problems would help to alleviate some of these risks. In this work, a suggested solution to help the physicians for fast pregnant diagnosis is proposed. A smart system is designed for easing women diagnosis during pregnancy. The core purpose of this research work is to build up an Android application which gives a deliberate approach in providing the best diagnosis of pregnancy so that women will not be caught in the is the complications during pregnancy. Android application uses Machine Learning algorithms such as decision tree algorithm. In this paper, decision tree algorithm (DT) is used for predicting pregnancy states, which performs better and more robust than the other models. The best diagnosis of pregnancy is predicted using C4.5 and CART algorithms. Proposed system had used decision tree algorithms that employed a database of the most diseases associated with women during pregnancy. Pregnant women is interacted with proposed system via an android application mobile interface, they had provided the algorithm with the case conditions, which processed to guide both physician and pregnant women on the level of risk. Response of the system is based on input data that processed according to predetermined variables and decision tree outcomes. Implemented experiments had proofed the efficiency of the system to facilitate diagnosis for better pregnant mothers' health care. The CART algorithm has gained a better level of evaluation comparing to other C4.5 algorithm. CART algorithm has achieved the highest correctly classified instances and incorrectly classified instances, which were 89.65% in correctly classified instances, 90.77 % in incorrectly classified instances, 9.23% evaluation measure.