Development of the method of extracting causal expressions from Japanese financial articles concerning business performance and analyzing them for supporting decision on investment

我们开发了一种从日本金融文章中提取有关公司经营业绩的因果表达式并自动分配其重要性的方法。分配重要性的因果表达式对投资者选择投资公司、预测业务和预测技术趋势很有用。该方法利用统计信息和初始线索表达式自动提取因果信息作为因果表达式的一种形式。我们从这些公司的网站中提取关键字，并利用这些关键字对之前的方法进行改进。此外，我们的方法通过使用从公司网站中提取的这些关键字来自动确定最重要的因果表达式。我们评估了我们的方法，它达到了77。实验结果表明，我们的方法优于Lead（s2）方法，该方法确定了一篇关于公司经营业绩的文章中出现的第2个句子作为最重要的因果表达，以及Lead（ce）方法，该方法确定了一篇关于公司经营业绩的文章中首先提取的因果表达作为最重要的因果表达。

We have developed a method of extracting causal expressions from Japanese financial articles concerning business performance of companies and assigning importance to them automatically. Causal expressions assigned importance are useful for investors in selecting companies to invest, business prediction, and technology-trends prediction. Our method automatically extracted causal information as a form of causal expression by using statistical information and initial clue expressions. We extracted keywords from Web sites of the companies and improved the previous method by using them. Moreover, our method automatically determined the most important causal expression by using these keywords extracted from Web sites of the companies.We evaluated our method and it attained 77. 3% accuracy and the experimental results showed that our method outperforms Lead(s2) method which determines the sentence which appears in the 2nd in an article concerning business performance of companies as the most important causal expression and Lead(ce) method which determines a causal expression first extracted from an article concerning business performance of companies as the most important causal expression.