The Extraction of Process Information from Chinese Text Document and to Construct Process Model

• Main Authors: Yu-Hsin Lin, 林予歆
• Other Authors: Chao Ou-Yang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/85705112095840357812

碩士 === 國立臺灣科技大學 === 工業管理系 === 96 === Due to the rapid development of enterprise globally, the documents used to describe business process have become very tedious and might be difficult to catch the required information. In addition, the process diagrams embedded in the documents usually were not properly mapped with the text description. This might due to either a very detailed text description versus a simplified process diagram or vise versa. However, both of the situations might mislead the readers due to incompatible information between text and process diagram. This research tends to propose a method to extract the process related information from Chinese text documents and to construct the related process diagram. An approach using word and sentence segmentation would be developed. The segmented words and sentences would be extracted and added tags. An algorithm will be developed to compute the precedence and logical relationships among the tagged words and sentences and to construct the related process models. The constructed models would be compared with the process diagrams embedded in the original document in terms of the index such as precision and recall. The computed values can be used as a reference to analyze the variance among the constructed and original process models. This research will be addressed on both of the English and Chinese documents in two years of period.