I need to translate Spanish tweets into english for my research. I find some toolkit. Among them, Moses is used by some research papers and other emerging toolkits used them as a baseline for evaluation purpose. So i am considering it as a candidate. Also, I found a toolkit from Stanford university called Phrsal, which also seems to be good. The last one I found is from renowned nltk
library. It also has a translate package. Every one of them states that they used phrase based statistical machine translation
technique along with some other techinques. Now my question is, from a practical or theoretical point of view, which will be best to use for tweets translation. Or google translator api
would be the best solution?
Asked
Active
Viewed 123 times
1

sovon
- 877
- 2
- 12
- 28
-
1You might as well take a look at [Google Neural Machine Translation](https://research.googleblog.com/2016/09/a-neural-network-for-machine.html). We're investigating whether we can use this for our company documentation. It isn't statistical anymore, but based on neural networks. You can [try it yourselves](https://github.com/tensorflow/tensorflow). – Nander Speerstra Jan 02 '17 at 14:50