一个基于金融混合数据的智能问答模型

智能问答系统（Question Answering System,QAS）是一种让人类通过自然语言与机器进行问答来获取信息的人机交互系统，是自然语言处理的一个集综合性与广泛性于一体的应用．现有的智能问答系统模型研究主要考虑单一型数据．然而，现实生活中结合表格和文本的混合型数据十分普遍，如金融领域的财务报表．本文对已有的财务报表智能问答模型进行改进并提出了一个新模型，该模型有更好的效果．

A question answering model based on hybrid data in finance domain

Question Answering System (QAS) allows humans to discuss questions with machine through natural language and is a comprehensive and extensive application of the method of natural language processing. Nowadays, a QAS is mainly based on the single-structured data and thus not suitable for the mixed data combining both table and text, such as the financial statements. This paper proposes a new QA model to deal with such questions automatically. Numerical simulations show that the new model has better performance.