Abstract Due to the sparsity of data and the limitation of bilingual data size, many high-quality phrase pairs can’t be generated. The example-based phrase pairs proposed by the authors are generated through decomposing, substituting and generating the typical phrase pairs, and the typical phrase pairs are generated by the typical phrase extraction method in phrase-based statistical machine translation. On the Chinese-to-English Newswire and Oral translation tasks, the experimental results demonstrate significant improvements achieved by the proposed methods. Moreover, a gain of about 1% BLEU score increase is yielded on some test sets.

The major error sources of deriving TEC from GPS observational data are instrumental biases. Two methods of extracting TEC and instrumental biases based on single site's GPS observations, called Kalman filter and self adjusted cell methods respectively, are presented. TEC and instrumental biases were computed by using these two methods from GPS data at several sites in 2004. The results were analyzed respectively and then compared with each other. The results also were compared with corresponding values published by CODE (Center for Orbit Determination in Europe).Its statistical characteristics with respect to the latitude of observational site were discussed. The main results can be concluded that the proposed methods can be used to derive TEC and instrumental biases which are accuracy enough with respect to CODE's published values. The standard deviations of derived instrumental biases are related to latitudes of observational sites. For China region, based on the paper's computational results, it shows that the sites with lower latitude are related with larger deviation, which is a implication that ionosphere is more complex in low latitude area than that in mid latitude and more caution should be paid in using the assumption of simple smooth ionospheric model in lower latitude.