Interactive Aided Translation Technology Research Based On Internet

Machine translation was studied for a long time, but hasn't been able to achieve the desired goal of mankind. With the rapid development of computer software and hardware technology, and the corpus for development, using the knowledge of statistical machine translation is expected to improve translation quality. Since the noise channel model, especially the maximum entropy model is put forward; machine translation of a central task is how to more effectively in the model of knowledge to further improve the quality of the machine translation. This article will focus on machine translation between Chinese and English, on how to combine with user evaluation effectively based on bilingual corpus in machine translation.Firstly, web crawler is an automatic extraction procedure, and we use it to download from the Internet web, get corresponding of Chinese English parallel corpus, stored in a database. Due to the Chinese for the basic writing on words, there is no clear distinction between the words mark. Therefore, Chinese words analyses are based on Chinese information processing and key. One Chinese word segmentation, we adopted the existing Chinese academy of sciences Institute of Computing Technology, developed from Chinese Lexical Analysis System of Technology Institute ICTCLAS (Institute of Computing Technology. Chinese Lexical Analysis System). But English sentence itself is separated already; we only had case transformation and punctuation separation treatment.Secondly, from the Internet to get back on the Chinese-English parallel corpora and processed, sentences for each translation of artificial score. Score for splendidly, good, general, poor, error 5 grades. The corpus of access and score screenshots appendix. Then, we will make different people get a score, and the average is multiplied by co-occurrence probability of statistical translation. The probability of the final translation probability will decode with Moses.Finally, we use the open source statistical machine translation of the Moses decoder, on the basis of the English language model joined trigger, combining the translation model and user ratings of tone sequence model, Chinese-English corpus appears in the data smoothing sparse phenomenon. Experimental results show that translation result is reasonable.