AUTOMATIC PATTERNS ACQUISITION AND EVALUATION FOR WEB-BASED TERMINOLOGY TRANSLATION
To find the translation of a given terminology from web without dictionary is an interesting and challengeable work.The existing methods based on pattern mainly consist of two steps: (1) learn patterns from a training set, and (2) find the candidate terms and score them.However, there are two main deficiencies in the existing works: (1) the amount and reliability of patterns are restricted by the training set, and (2) the methods for scoring patterns and candidate terminologies are too simplified.We present a new method that needs only a pair of terminologies for training to acquire the initial patterns.Patterns acquisition and patterns evaluation are performed automatically when an appropriate candidate terminology is selected for a users query.We also improve the method of scoring the candidate terminologies by applying heuristic rules.The experiment results show that our method is better than the existing technologies.
Pattern acquisition Automatic evaluation Web-based Terminology translation
ZHI-SHENG LI PI-LIAN HE YUE-HENG SUN
Tianjin University, College of Computer Science and Technology, Tianjin, 300072, China
国际会议
2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)
香港
英文
4124-4129
2007-08-19(万方平台首次上网日期,不代表论文的发表时间)