Article
Algorithm for Arabic-English Hybrid Machine Translation Optimisation
The distinct designs of Rule Based Machine Translation and Statistical Machine Translation make it difficult for a pure machine translation solution to handle all translation obstacles. Once lexical and syntactic analyzers are dealt with using Rule Based, the Expectation–Maximization algorithm is run to resolve certain uncertainties with maximum likelihood. We have designed and explained an integrated hybrid machine translation system in this paper. Fusing the most effective features from different strategies is what is pursued. In the first step, RBMT converts Arabic text to English and then EM is run on the output to improve it and create the final version. To do well and become better, the EM algorithm must be able to accurately translate a frequency from one language to another. The results prove that the new approach works better than the common Rule Based approach and the EM algorithm, as proven by the BLEU system. The survey suggests that scores from the HMT system are regularly greater than those from SMT.
Full Text Attachment
