A Study of automatic tallying-up method for open-ended questions in the questionnaire survey and automatic generation for select appropriate questions

问卷调查中开放式问题的自动统计方法及自动生成选题的研究

• 批准号: 14580350
• 负责人: YAGUCHI Hiroyuki
• 金额: $ 0.9万
• 依托单位: Tokyo Denki University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2002
• 资助国家: 日本
• 起止时间: 2002 至 2004
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/en/grant/KAKENHI-PROJECT-14580350/
• 关键词: text mining; questionnaire surveys; open-ended questions; Morphological analysis; Clustering; 自由回答文; 自動集計; 自由解答; 自由回答

The First year of our research, we developed the sentence clustering systems which classify the sentences in questionnaire survey. Our sentence clustering system classifies sentences based on the type of morpheme and it's kind of the parts of speech which form the sentences. We questionnaire survey did in order to gather the data which is necessary for the test of our system. We examined 3 methods for suitable data entry method from OCR, speech recognition and typing. The result of experiment, we find the most suited data entry method as speech recognition.2nd year of our research, we examine our system in use some Questionnaire surveys. The result of experiment, we confirm our developed system is effective.The last year of our research, mainly we appraised our system between our system and other system. We research commercial based text mining systems, and we compeer with our system with another text mining system. At the results, we find our system have same performance with another commercial base text mining system. And we find week point of our system. In the future, we continue development of our system.

我们研究的第一年，我们开发了句子聚类系统，对问卷调查中的句子进行分类。我们的句子聚类系统根据语素类型和构成句子的词类对句子进行分类。我们进行问卷调查是为了收集测试我们的系统所需的数据。我们研究了 OCR、语音识别和打字等 3 种合适的数据输入方法。实验的结果，我们找到了最适合的数据输入方法作为语音识别。我们研究的第二年，我们使用一些问卷调查来检查我们的系统。实验的结果，我们证实了我们开发的系统是有效的。去年我们的研究，主要是我们的系统与其他系统之间的评价。我们研究基于商业的文本挖掘系统，并将我们的系统与另一个文本挖掘系统进行比较。结果，我们发现我们的系统与另一个商业基础文本挖掘系统具有相同的性能。我们找到系统的周点。未来，我们将继续开发我们的系统。