In today's global marketplace, translators are responding to the current demand to produce fast and high-quality translations by using electronic tools to help them do their jobs, and one of the most promising tools that translators have at their disposal is translation memory (TM) software, a veritable database of previously translated material. Fuzzy matching---where the TM system identifies a portion of the text that is similar to but not exactly the same as one stored in the database---has become an integral feature of TM software; yet using this feature effectively remains a mystery to most translators. This is largely because translators have not been presented with any type of guidelines with regard to helping them identify an ideal setting for the fuzzy match value.
The objective of this thesis is to provide translators with a better understanding of TM systems by exploring fuzzy matching in detail, and particularly by investigating factors such as the TM system selected, the category of text being processed, the working languages involved, and the degree of fuzziness of the match. Accordingly, a series of experiments have been designed and carried out to determine the influence that these four factors might have on the ideal fuzzy match setting. The results of these experiments show that these factors should indeed be taken into account when translators are selecting the fuzzy match value to be used with a TM system.
Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/27516 |
Date | January 2007 |
Creators | Fifer, Matthew |
Publisher | University of Ottawa (Canada) |
Source Sets | Université d’Ottawa |
Language | English |
Detected Language | English |
Type | Thesis |
Format | 174 p. |
Page generated in 0.0016 seconds