A Smart Advisor for Pregnancy Healthcare Using Chatbot Approach Based on Decision Tree Methods

Abstract

With the technological progress in the current era, artificial intelligence has gained significant popularity in all applicable fields. Indeed, one of these important fields is automated chat systems or so-called smart conversation chatbots. In this study, we introduce a mobile application for automated chat with pregnant women. We have employed this technology in providing a healthcare assistant. As bringing a new life into the world is the wish of every mother, it proposed health care system helps the pregnant mother by discussing some symptoms with the pregnant mother, diagnosing diseases, and predicting the mode of delivery. It is used for providing guidance regarding nutrition, healthy exercises, safe medications, and many different functions pregnant mother needs. The proposed chatbot system is not based only on rules, rather, it has been built based on the machine learning (ML) method; i.e., decision trees (DT) model. ML methods help process user inputs, which are not previously entered during training, and then answer them in a good level of correctness. Besides, the system overcomes most spelling and linguistic errors, accepts words with diacritics, and many other features. After training the smart advisor several times, it was tested by introducing several different conversational formulas. The evaluation accuracy results reached 85.45%.