Takeno et al. \citeyear{takeno:17} adapted the method for English-Japanese translation.
To date, no comparisons of the performance impact between the prefix constraints have been made.
Thus the question arises if the use of prefix constrains mechanisms can be generalized and used with any language pair.
+
+
+## Aims
+The problem of domain control is not yet solved and current research.
+In this thesis I want to evaluate the domain control mechanism prefix contraint and how it impacts the translation performance in different language pairs.
+I want to analyze additonaly if the relatedness of a language pair needs to be considered during multi domain corpus translation.
\chapter{Conclusion}
In conclusion, the problem of domain control in neural machine translation is very challenging and not solved in any translation-related area.
-The impact of the domain control mechanism can not always be predicted and depends on multiple factors (e.g. model, domain, language, etc.).
+The impact of the domain control mechanism prefix constraints can not always be predicted and depends on the similarity of domain distinction in the languages.
The domain and language selection should always be considered when choosing any domain adaption mechanism.
+The connection between the impact of prefix constraints and the relatedness needs to be researched further.