RECall: A Scheduling System and Question Answering System with User Knowledge Base on a Mobile Application for Remembering and Recovering information

This paper provides a tool that aids users for remembering and recovering information on a mobile application, named RECall. The user can record his knowledge like schedules, medications or other information through the app and he can retrieve them by asking questions. Question answering (QA) aims to answer a natural language question. The challenge, however, is how to deal with the same questions that have different ways of expressing it. Knowledge-based question answering makes use of a knowledge base (KB) as an information source where the data is structured. Structuring the knowledge representation will be essential for generating answers later on, which is done by information extraction (IE), for a knowledge base containing logs, schedules, and medications of the user, for this study. The question answering focuses an open domain, and it uses keyword-based, synonym-based, and rule-based approaches. While, scheduling (SCH) allows adding schedule, time-based moving schedule, entity-based or time-based canceling schedule, and recording medication with calendar app integration. Given a dataset of 100 knowledge and 100 questions, the question answering evaluated about 79% accuracy for classifying correct, incorrect, and null answers out of 95 valid questions, with 53% and above precision, recall, and F-measure. Although there are more rules constructed upon resolving rules in conflict, with a solid foundation and combining other approaches, it leaves the possibility of being open to more cases. Nevertheless, the study has achieved its purpose to develop a mobile application for scheduling and question answering systems with Representational State Transfer (REST) web services.