Question answering (QA) is a system that can automatically answer questions posed by humans. Currently, QA system finds answers that are already known by somebody and have been written in some web pages, rather than generating new knowledges.

My QA system was still under development as I was writing the report. Now it can only extract keywords and define the answer type. Extracting answers part will be the tasks for next semester.

There are different types of questions and answers. My system design focuses on the answer type. My system will first extract key words of the question to see what type of answer it should generate. The answer types can be put into these classes:

• Person (from "Who . . . ")
• Place (from "Where . . . ")
• Date (from "When . . . ")
• Number (from "How many . . . ")
• Explanation (from "Why . . . ")
• Method (from "How . . . ")

Then I used POS tagging to tag the original question. Nouns and verbs are the primary key words in the question and other words are the secondary keywords. Stop words should also be removed. Now I have obtained the keywords for the question. The following steps are the task for me to do in the future implementation.

After that, I need to use the keywords to generate query for Yioop Query System to find search related pages for me to get candidate answers.

My system will go through the candidate answers and extract sentences from it. The Chinese sentences are different from the English sentences since one Chinese sentence might contain multiple English sentences. You can think about some English sentences separated by comma instead of period. That is caused by Chinese gramma. In Chinese, if one sentence has some relationship with the sentence after it, you can use a comma instead of a period and you can even omit the subject of the sentence. That causes some problems when I was trying to extract answers. My next plan is to deal with these problems and complete the system.