How-Q Visualized answer Integrated causality graph A B S T R A C T The research aims to develop an automatic Question Answering system, in particular Why and How questions, on community web-boards to support ordinary people in preliminary diagnosis and problem solving, such as plant disease problems. The research includes two main problems: Why and How question identification and Why and How answer determination, where Why and How questions are based on explanations. Therefore, the research

more »

... ies machine learning techniques for question type identification. We also propose an integrated causality graph with extracted procedural knowledge from text to determine the visualized answers based on the information retrieval technique. The experiment shows the Question Answering system can achieve answers at Rank 1 with 91.1% and 88.9% correctness for Why questions and How questions, respectively.